Portal de Periódicos da CAPES

The Renal Safety of Sevoflurane

2000; Lippincott Williams & Wilkins; Volume: 90; Issue: 3 Linguagem: Inglês

10.1097/00000539-200003000-00001

1526-7598

Robert F. Bedford, Harlan E. Ives,

Cardiac, Anesthesia and Surgical Outcomes

Since its introduction into American clinical practice in 1995, sevoflurane has been given to tens of millions of patients without a single report of nephrotoxicity, either in the scientific literature or to the Food and Drug Administration (FDA). Accordingly, it might seem curious that the current labeling for sevoflurane continues to warn against its prolonged administration at relatively low fresh gas flow rates (≤1 L/min) because of the potential risks of patient exposure to Compound A [CH2F–O–C (=CF2)(CF3)], one of the degradation products of sevoflurane [CH2F–O–CH (CF3)2] resulting from its interaction with standard carbon dioxide (CO2) absorbents. Indeed, the article by Mazze et al. (1) in the current issue appears to call into question the necessity of a fresh gas flow rate warning in the prescribing information for sevoflurane. Among 1,941 surgical patients given sevoflurane, Mazze et al. (1) found that sevoflurane did not influence serum creatinine differently from comparative anesthetics (primarily isoflurane) given to 1,495 patients. The large sample size also enabled tests of the effect of pre-existing abnormalities in serum creatinine and of administration of potentially nephrotoxic antibiotics, and these studies also indicated that the effects of sevoflurane did not differ from those of comparative anesthetics: no anesthetic was associated with a clinically significant increase in serum creatinine. Thus, the report by Mazze et al. (1) supports the view that sevoflurane is a safe anesthetic that does not adversely affect the kidney when used as presently recommended in the package insert. Given the clinical advantages offered by sevoflurane, particularly an absence of pungency and a relatively low solubility, such a sense of safety is welcome. Why, then, the FDA’s insistence on a minimal fresh gas flow rate warning? The answer lies in three limitations to the Mazze et al. (1) study. First, 91% of the patients received sevoflurane at fresh gas inflow rates of >2 L/min. Such high inflow rates limit rebreathing and thereby minimize the respired concentration of Compound A because Compound A arises from the action of CO2 absorbents on sevoflurane. Thus, Mazze et al. (1) did not rigorously test the capacity of Compound A to produce renal injury. Second, 97% of Mazze et al.’s (1) patients received less than 4 minimum alveolar anesthetic concentration-(MAC) h of anesthesia and, thus, do not provide a test of the effect of prolonged sevoflurane anesthesia. However, others have tested the effect of prolonged sevoflurane anesthesia, albeit in far smaller numbers than examined by Mazze et al. (1). All such tests find that prolonged (>5 MAC-h) sevoflurane anesthesia does not increase serum creatinine (2–5). A third and critical limitation of the Mazze et al. (1) study is the reliance on serum creatinine level as the primary marker of renal insult. Serum creatinine concentration is a powerful tool that reflects glomerular filtration rate. However, although glomerular filtration rate declines (serum creatinine rises) in certain renal diseases, many diseases damage the kidney without decreasing glomerular filtration rate (6). These include glomerular and tubulointerstitial diseases. Thus, serum creatinine is a good marker of neither increased glomerular permeability (“leakiness”) nor tubular integrity, and despite the power of serum creatinine as a tool, the thorough physician would be remiss in using creatinine as the sole measure of renal integrity. Markers of increased glomerular permeability include proteinuria, particularly albuminuria. Markers of tubular integrity include glucosuria and enzymuria [e.g., the appearance of abnormal levels of N-acetyl-β-glucoseaminidase (NAG) or α-glutathione-S-transferase in the urine]. Mazze et al. (1) argue that only increased serum creatinine and, perhaps albuminuria, are validated markers of renal disease, and thus, the use of other markers is unwarranted. However, validation (i.e., substantiation by histopathology) has been documented primarily for disease rather than for drug toxicity, and toxic drugs may target tubules rather than the glomerulus. As indicated in the preceding paragraph, reliance on serum creatinine neglects injury to the glomerulus that allows increased permeability of albumin but does not affect glomerular filtration capacity and neglects injury that occurs to the tubule rather than the glomerulus. Tubular injury is important in the context of sevoflurane safety because Compound A can affect the renal tubule (3,6). In rats, Compound A can produce renal tubular necrosis (of the corticomedullary junction) (7–10). Tubular necrosis appears at lower doses of Compound A (defined as Compound A concentration multiplied by duration of exposure) than those associated with increases in serum creatinine (9). Doses <50–180 ppm-h of Compound A do not cause histological evidence of necrosis, and increasing doses of Compound A above this “threshold” increase necrosis (Table 1) (7–10). Doses of Compound A associated with histological evidence of tubular necrosis also are associated with proteinuria and enzymuria (9). Similarly, Cynomolgus monkeys given 100 ppm Compound A for 8 h have increased urinary protein and urinary NAG in association with single cell tubular necrosis and tubular degeneration (11).Table 1: Studies Examining Renal Insult After Compound A ExposureIn humans, several investigations report an association of Compound A exposure and albuminuria, glucosuria, and/or enzymuria (increased urinary α-glutathione-S-transferase or NAG) (2–5,12), and several investigations do not (12–15). The finding of albuminuria, glucosuria, and/or enzymuria appears to be associated with inhalation of doses of Compound A exceeding 160 ppm-h: For example, albuminuria (or an increase in urinary albumin excretion) is not found with doses 160 ppm-h may be associated with albuminuria (Table 1) (2–5,12). In all studies, the albuminuria, glucosuria, and enzymuria associated with larger doses of Compound A has been transient, lasting 3 to 5 days with return to normal values usually in less than one wk. In no study does serum creatinine correlate with albuminuria, enzymuria or glucosuria. The finding of albuminuria, glucosuria, and/or enzymuria does not appear to result from anesthesia, per se, or from sevoflurane when exposure to Compound A is minimized. Among 14 patients given sevoflurane for 6.7 h (11.0 MAC-h) at a 6 L/min flow rate (approximately 30 ppm-h of Compound A exposure), no patient had proteinuria or glucosuria (4). In contrast, 10 of 14 patients given sevoflurane for 6.7 h (10.9 MAC-h) at a 1 L/min flow rate (192 ppm-h of Compound A exposure) had proteinuria, and 3 had glucosuria. In the same study, none of 14 patients given isoflurane for 6.7 h (10.9 MAC-h) had glucosuria and 1 had proteinuria. In another study, administration of 1.25 MAC sevoflurane to volunteers for 8 h (10 MAC-h) at a 2 L/min inflow rate (Compound A dose = 348 ppm-h) was associated with albuminuria, glucosuria, and enzymuria (5). When the same volunteers were given 1.25 MAC desflurane for 8 h, none had albuminuria, glucosuria, or enzymuria. Thus, studies in rats, monkeys, and humans indicate that increasing doses of Compound A above a threshold (approximately 160 ppm-h in rats and humans; approximately 5 times greater in monkeys) produce a dose-related albuminuria, glucosuria, and enzymuria. In rats and monkeys, there is an associated dose-related increase in histological evidence of renal tubular necrosis. No studies have assessed this association in humans because of the risks imposed by renal biopsy. Thus, unanswered is the question of whether the transient albuminuria, glucosuria, and enzymuria found in humans indicate transient renal necrosis. How do these abnormalities compare with nonpathological (functional) changes in renal function and integrity? Functional changes usually are neither as striking nor as long lasting as those associated with higher doses of Compound A. For example, fever may produce functional albuminuria with median values of 22–28 mg urinary albumin per 24 h (16), but higher doses of Compound A can produce 1,000–4,000 mg per 24 h, and this albuminuria can last for days after elimination of Compound A. Prolonged vigorous exercise also may increase urinary albumin excretion. Albumin in the urine slightly increases after a 100-km hill walk (17), and 30 min after running a marathon (a period of maximal albumin excretion), albumin in the urine may increase 10 fold (18). If continued for 24 h, this level of albuminuria would produce 180 mg albumin in the urine (i.e., much less than has been produced by exposure to Compound A). Furthermore, the increase with vigorous exercise is not sustained, having a half-life of 54 min (18). Thus, it appears unlikely that functional changes explain the albuminuria, glucosuria, and enzymuria seen in humans after exposure to higher concentrations of Compound A. In summary, the available information indicates that all durations and depths of sevoflurane anesthesia are nontoxic to the normal human kidney as long as exposure to Compound A is kept below <150 ppm-h. Significant questions persist, however, regarding the potential for Compound A to cause renal injury at larger doses and whether transient albuminuria, glucosuria, and/or enzymuria reflect pathological changes (renal injury) or functional abnormalities. Does the usually mild and transient nature of the changes indicate that they are not clinically relevant? Need the practitioner consider the possibility that insults to the kidney may have a cumulative effect that may not be apparent in one or two exposures? Are particular patients more vulnerable to greater degrees of injury? Answers to these questions await further investigation. And what conditions predispose to increased Compound A concentrations? In a circle absorber anesthesia circuit, Compound A concentrations correlate directly with sevoflurane concentrations and absorbent temperature and inversely with fresh gas inflow rate (19). Increasing inflow rate decreases Compound A concentrations by decreasing rebreathing of gas emanating from the absorbent and by decreasing the amount of CO2 that reaches the absorbent (the amount of CO2 determines the temperature of the absorbent) (19,20). At inflow rates of 1 L/min, a 1 MAC end-tidal concentration of sevoflurane can produce 25–35 ppm of Compound A; 1.25 to 1.5 MAC produce proportionately more (5,21). The study by Mazze et al. (1) notwithstanding, it is because of the possibility that Compound A may cause renal injury that the Warnings section of the latest sevoflurane package insert continues to recommend that Compound A exposure be minimized by limiting the use of higher concentrations of sevoflurane at low inflow rates for longer periods of time: “Although data from controlled clinical studies at low flow rates are limited, findings taken from patient and animal studies suggest that there is a potential for renal injury which is presumed due to compound A. Animal and human studies demonstrate that sevoflurane administered for more than 2 MAC-hours and at fresh gas flow rates of <2 L/minute may be associated with proteinuria and glycosuria. While a level of compound A exposure at which clinical nephrotoxicity might be expected to occur has not been established, it is prudent to consider all of the factors leading to compound A exposure in humans, especially duration of exposure, fresh gas flow rate, and concentration of sevoflurane. During sevoflurane anesthesia, the clinician should adjust inspired concentration and fresh gas flow rates to minimize exposure to compound A. To minimize exposure to compound A, sevoflurane exposure should not exceed 2 MAC-hours at flow rates of 1 to <2 L/min. Fresh gas flow rates <1 L/minute are not recommended.” Whether this statement is prudent or excessively conservative remains to be determined. Finally, concerns regarding the degradation of anesthetics by CO2 absorbents to toxic compounds may become moot. The widespread application of absorbents that minimally degrade sevoflurane to Compound A or desflurane to carbon monoxide would eliminate any potential hazard from these toxic compounds (22–26). Such absorbents do not contain either sodium hydroxide or potassium hydroxide, both of which appear to enhance the production of Compound A and carbon monoxide. Examples of such absorbents include calcium hydroxide as in Amsorb (Armstrong Medica Ltd. Coleraine, Northern Ireland) or lithium hydroxide (Dornier GmBH, Friedrichshafen, Germany). Both Amsorb and lithium hydroxide are used commercially as carbon dioxide absorbents. Amsorb is used in anesthetic circuits in Europe and lithium hydroxide is used in aerospace and underwater equipment. Both absorbents decrease production of Compound A and carbon monoxide to clinically insignificant levels (personal communication from Dr. Caroline R. Stabernack, University of California, San Francisco, CA 94143-0464). The authors wish to acknowledge the suggestions made by Edmond I Eger, II, MD, in the development of this manuscript. Dr. Eger is a paid consultant to Baxter PPI, the manufacturer of desflurane.