A Case of Nonoliguric Renal Failure After General Anesthesia with Sevoflurane and Desflurane

Biodegradation of sevoflurane to fluoride ion has raised concerns regarding the nephrotoxicity of sevoflurane in humans [1,2]. In addition, soda lime and Baralyme[registered sign] degrade sevoflurane to a nephrotoxin (pentafluoroisopropenyl fluoromethyl ether) known as Compound A [3-7]. Although sevofluranerelated nephrotoxicity has been demonstrated in volunteers [8], nephrotoxicity severe enough to require therapeutic action has not been reported. We report acute nonoliguric renal failure occurring in a 25-year-old man with previously normal renal function after anesthesia with sevoflurane and desflurane. Case Report A 25-yr-old, 52-kg man was admitted to our burn unit with 20% partial and full-thickness flame burns on his neck and both arms. He had no prior surgeries or allergies and was not taking medication. He denied the use of alcohol or illicit drugs. On admission, his vital signs were normal, and his physical examination was otherwise unremarkable. Initial laboratory results were normal (Table 1). The Parkland formula for burn resuscitation (4 mL/kg per percentage of burn over 24 hr) guided initial intravenous (IV) fluid management [9].Table 1: Chronological Events and Laboratory FindingsaHis subsequent hospital course extended over 39 days and included four general anesthetics (Table 1). Anesthesia was induced and maintained with thiopental and isoflurane in the first two cases. Sevoflurane at a 0.3% average end-tidal concentration was used for 20 min at the end of the second case. In the third case, sevoflurane at a 1.0% average endtidal concentration was used for 60 min after the induction of anesthesia with propofol. In the fourth case, after induction with thiopental, sevoflurane at a 1.7% average end-tidal concentration was used for 45 min, and desflurane was used for 2 h at low flows to conclude the anesthetic. Neither hypotension nor decreased urine output occurred at any time during hospitalization, and pre- and postoperative blood urea nitrogen (BUN) and creatinine for the first three surgeries were normal. On Hospital Day 4, haloperidol (2 mg IV every night) was given to treat signs of paranoid ideation. Paroxetine (10 mg per os every morning) was added on Day 10 to treat overt depression. Hyponatremia, probably secondary to paroxetine, developed over the next 4 days (Na+ = 122 mEq/L) and was treated initially with oral fluid restriction and subsequently with IV saline. On Day 11, penicillin G (1.5 million U IV every 4 h) was given to treat cellulitis associated with temperatures ranging from 38.4[degree sign]C to 39.1[degree sign]C. Hemodynamic variables remained stable, and laboratory variables continued unchanged. On Day 12, penicillin was changed to vancomycin (1 g IV bid) to reflect organism sensitivities. On Day 15, a vancomycin peak level of 17.9 micro g/mL led to an increase in dose (1.25 g IV bid), and haloperidol was discontinued. On Day 16, a fourth anesthetic was given for left arm debridement. Sevoflurane (1.7% average end-tidal concentration) was administered in a 2-L/min inflow containing 50% nitrous oxide. After 45 min, at the preference of the attending anesthesiologist, sevoflurane was discontinued and 8% inspired desflurane was begun in a total gas flow of 0.55 L/min. High flows were not used to eliminate residual sevoflurane. The procedure lasted another 2 h. Urine output for the first hour was 80 mL and declined to 37 mL the second hour. Tachycardia (i.e., 115 bpm), normally observed in burn patients and consistent with this patient's preoperative baseline, was noted intra- and postoperatively. Arterial blood pressure remained at 110-80/51-75 mm Hg. Esophageal temperature was 37.5 [degree sign] C, and room temperature ranged from 30 [degree sign] C to 35 [degree sign] C. No IV contrast agents, topical antibiotics, nonsteroidal antiinflammatory drugs, or sympathomimetics were given. Crystalloid, 1400 mL, and hetastarch, 1000 mL, were infused. Estimated blood loss was 300 mL. Over the next 48 h, BUN and creatinine levels progressively increased to a peak of 40 mg/dL (normal 7-20 mg/dL) and 2.2 mg/dL (normal 0.5-1.4 mg/dL), respectively, although no adverse hemodynamic events had occurred and urine output remained unchanged in volume and color. On Day 17 (Postoperative Day 1), urine protein was 0.48 g/d (normal 0.05-0.10 g/d). Vancomycin administration was discontinued in expectation of continued renal failure. On Day 18, a dipstick measurement revealed 1 + blood and no glucose. Urine specific gravity was 1.011, urine pH was 5, and the urine creatinine level was 46 mg/dL. Moderate hyaline and granular casts were found. Urine eosinophils were absent and serum eosinophils were normal. Serum creatine kinase and lactate dehydrogenase levels were normal. Fractional excretion of Na+ was 3%. Renal function returned to baseline by Day 24. The patient was discharged in good condition on Hospital Day 39. Discussion We present an unusual case of acute nonoliguric renal failure after routine surgery. No clear cause could be identified from the patient's record. Known risk factors for perioperative renal failure [10], such as hypotension, preexisting renal disease, or left ventricular dysfunction, were not present. Additionally, most cases of postoperative renal failure are oliguric in nature [11]. Nonoliguric renal failure does occur, but it is usually associated with the administration of nephrotoxins such as aminoglycosides [12]. The pattern of renal failure was unusual. The rapid increase and decrease in the serum creatinine level suggests an acute, transient, 75% decrement in creatinine clearance over a period of 48 hours without change in urine output, intravascular volume status, hemodynamics or clinical condition. Hypovolemia leading to renal hypoperfusion cannot easily explain this pattern of renal failure. Positive fluid balances were measured over the three days before the procedure, and perioperative urine output exceeded 2000 mL/d. Although a preexisting concentrating defect could lead to large urine output in the presence of hypovolemia, previously obtained urine electrolytes (data not shown) clearly demonstrated this patient's urine concentrating ability. Systemic sepsis is also an unlikely explanation for this event. Acute renal failure from sepsis is usually oliguric, and patients are often hemodynamically unstable. Additionally, the episode of renal failure resolved in the continued presence of fever. Vancomycin alone is an improbable cause of this patient's renal failure. Cases of vancomycin-induced nephrotoxicity are now thought to be caused by impurities in older preparations of the drug [13]. In addition, most reports found vancomycin nephrotoxicity only when vancomycin trough levels exceed 30 micro g/mL. Our patient had a vancomycin peak level of 17.9 micro g/mL 24 hours before surgery, and a random level of 28 micro g/mL was obtained postoperatively during the period of renal failure (three hours after the last dose). Current vancomycin preparations are not considered nephrotoxic, although the combination of vancomycin and aminoglycosides can cause synergistic toxicity [14]. Sevoflurane biotransformation to inorganic fluoride [1] and CO2 absorbent-catalyzed breakdown of sevoflurane to Compound A [3] have raised concerns about potential nephrotoxicity. Methoxyflurane, an older anesthetic also metabolized to fluoride, can produce polyuric renal failure with increased BUN and creatinine levels [15]. Healthy patients given sevoflurane may have inorganic fluoride levels equal to those associated with methoxyflurane toxicity but only minimal evidence of renal damage [1]. Compound A can produce renal injury with characteristic proteinuria in rats [6] and humans [8], a finding noted in our patient. Although our search failed to reveal a more likely cause, we think it unlikely that sevoflurane was the sole cause of this episode of renal failure. Our patient received approximately 0.8 minimum alveolar anesthetic concentration (MAC) hour of sevoflurane alone, plus an unknown additional exposure associated with rebreathing residual sevoflurane at low (0.55 L/min) flow rates. Although the package insert for sevoflurane (Abbott Laboratories, Chicago, IL) advises against administration at flow rates less than 2 L/min, there are no reports that document harm from this practice in patients. Both the increased room temperature (30[degree sign]C-35[degree sign]C) and a hypercatabolic state resulting from thermal injury might have increased sevoflurane degradation to Compound A by increasing the temperature of the soda lime. Exposure to sevoflurane for even three hours, however, is not usually associated with toxic fluoride or Compound A levels [16]. Could a synergistic interaction among vancomycin, sevoflurane, fever, and desflurane have caused this patient's acute renal failure? Although vancomycin itself is not nephrotoxic, it can potentiate aminoglycoside toxicity. Synergism among aminoglycosides, renal hypoperfusion, endotoxemia, and fever occurs in vitro and in animal models [17]. Such synergism may also exist between vancomycin and fluoride ion or Compound A. Many early reports of methoxyflurane toxicity, for example, occurred with the coadministration of methoxyflurane and tetracycline [18]. This patient also received desflurane for two hours during the fourth anesthetic. Desflurane biodegradation does not produce toxic fluoride levels, and no other toxic metabolites are known. Desflurane can transiently increase sympathetic activity when given at concentrations exceeding 1 MAC [19], possibly affecting renal blood flow. However, signs of increased sympathetic activity (increases in blood pressure or heart rate) did not follow the introduction of desflurane in this patient. Although desflurane alone is not considered nephrotoxic, concurrent exposure to both sevoflurane and desflurane may produce a synergistic toxicity by some unknown mechanism. Finally, we note that nephrotoxicity did not develop after the third anesthetic with sevoflurane and concurrent vancomycin administration. However, the doses of both sevoflurane (one hour at 1.0% average concentration) and vancomycin (1 g IV Bid) were smaller than those in the fourth anesthetic. In summary, we report a case of nonoliguric renal failure after an uncomplicated anesthetic with sevoflurane and desflurane. Had creatinine and urinary protein level measurements not been made, renal failure would not have been detected. No obvious cause for renal failure was found. Did sevoflurane contribute to this patient's renal failure, either directly or through some interaction with vancomycin, fever, or desflurane? Sevoflurane can cause similar renal injury in volunteers [8], but such injury occurs after longer exposure to higher sevoflurane concentrations. Although the insult was short-lived in this patient, other patients at greater risk of perioperative renal failure may be more susceptible to prolonged renal complications. We gratefully acknowledge the suggestions made by Abbott Laboratories and Dr. Edmond I Eger II.