Effects of Sevoflurane and Isoflurane on Renal Function and on Possible Markers of Nephrotoxicity 

Forty-two patients classified as American Society of Anesthesiologists physical status 1 (40 men, 2 women) who were scheduled to undergo dental or orthopedic surgery that was expected to last at least 4 h were studied. Patients who showed evidence of abnormal hepatic or renal function, based on medical history, physical examination, or laboratory tests, were excluded from the study. The hospital ethics committee approved the study, and patients gave written informed consent. Patients who were scheduled for surgery at each anatomic site (knee, shoulder, hand, and mandible; a tourniquet was inflated during the operation when the surgical site was in an extremity) were assigned consecutively to one of three groups (n = 14 in each group): the isoflurane group (anesthetized with isoflurane at a total flow of 1 l/min), the low-flow sevoflurane group (anesthetized with sevoflurane at 1 l/min), and the high-flow sevoflurane group (anesthetized with sevoflurane at 6 l/min). Isoflurane was selected as the control anesthetic because it undergoes less biotransformation and degradation by standard carbon dioxide adsorbents. [21,22]The high-flow sevoflurane group was included as a second control group because the use of higher fresh gas flow rates decreases the inspired concentration of compound A.

The anesthetic protocol was designed to result in prolonged high compound A concentrations. To increase the period of anesthesia, it was generally induced 60–90 min before the surgical procedure was expected to begin. Thirty minutes after receiving an intramuscular injection of atropine (0.5 mg) and midazolam (0.08 mg/kg), each patient received an intravenous injection of thiopental (3 - 5 mg/kg) and succinylcholine (1 mg/kg) or vecuronium bromide (0.1 mg/kg) to facilitate tracheal intubation. After tracheal intubation, anesthesia was maintained with isoflurane or sevoflurane, air, and oxygen (FiO²= 0.4) at a total flow rate of 6 l/min. After 5 min, the fresh gas flow rate was reduced to 1 l/min in the low-flow groups. A semiclosed-circle system with a soda lime (Dragersorb 800, Drager, Luebeck, Germany) was used to absorb carbon dioxide. The carbon dioxide absorbent was changed before the anesthetic was administered to each patient. The anesthesia machine was a North American Drager Narcomed IIB (Telford, PA). Anesthetic was administered via a Penlon PPV Sigma vaporizer (Penlon, Abingdon, United Kingdom) or a Muraco Forawic vaporizer (Muraco Medical, Tokyo, Japan). Two sevoflurane or isoflurane vaporizers were linked in series, permitting the administration of a high concentration of sevoflurane or isoflurane to patients in the low-flow system. The flow meters in the anesthesia machine were calibrated using Calibration Analyzer RT-200 (Allied Healthcare, St. Louis, MO) before each study. A radial arterial catheter was inserted to monitor arterial blood pressure and to obtain blood samples for analysis of arterial blood gases and serum inorganic fluoride concentrations. The lungs were ventilated mechanically to a tidal volume of 8 - 10 ml/kg, with the ventilatory rate adjusted to maintain an end-tidal carbon dioxide partial pressure of 35 - 40 mmHg. We connected an “artificial nose” to the endotracheal tube for airway humidification. End-tidal concentrations of sevoflurane or isoflurane were analyzed using a Capnomac Ultima gas analyzer (Capnomac, Datex, Finland) that was calibrated immediately before each study using a cylinder that contained a mixture of gases of known concentrations. Minimum alveolar concentration-hours for sevoflurane and isoflurane exposures were calculated from the percentage anesthetic concentration and the duration of anesthetic exposure. Minimum alveolar concentration values were 2.4% for sevoflurane and 1.28% for isoflurane for the age group studied. [23,24]Anesthetic concentration was adjusted by the anesthesiologist to maintain the mean arterial blood pressure within +/- 20% of baseline. No adjunct anesthetics nor vasoactive drugs were used. A temperature probe (temperature probe model DT-300, Intermedical Co., Tokyo, Japan) was inserted into the center of the upper absorbent canister, and soda lime temperature was recorded at 5-min intervals. The room temperature was maintained at 25 [degree sign]C, and the humidity was maintained at 50%. Anesthesia was maintained for at least 240 min, even if surgery was completed earlier than anticipated. After completion of the surgical procedure, anesthetic administration was discontinued, and the fresh gas inflow rate was changed to 6 l/min of oxygen. After the patients opened their eyes and took a deep breath after verbal command, the endotracheal tube was removed. Lactated Ringer's solution was administered at 5 - 6 ml [middle dot] kg^-1[middle dot] h^-1during anesthesia and at 2 ml [middle dot] kg^-1[middle dot] h^-1for 16 h after cessation of anesthetic exposure.

All patients received cefotiam as an antibiotic perioperatively, which was given intravenously twice a day (2 mg per day) from immediately after the induction of anesthesia to day 2 after anesthesia. Thereafter, 600 mg cefotiam was administered orally for 5 days.

Clinical laboratory measures of BUN, creatinine, serum aspartate aminotransferase, alanine aminotransferase, and glucose concentrations were performed immediately before anesthesia and repeated 1, 2, 3, 5, and 7 days after initiation of anesthesia. Urine samples (24 h) were collected before anesthesia and were continued until at least 7 days after anesthesia. These samples were used to measure urinary excretion of NAG, [small beta, Greek]^{2-microglobulin}, protein, glucose, creatinine, and inorganic fluoride concentrations. Postanesthetic urine collection began at the end of anesthesia for each 24-h period from 0 - 168 h. If urinalysis results were abnormal on day 7 after anesthesia (NAG, [small beta, Greek]^{2-microglobulin}, protein, and glucose), 24-h urine collection was continued until urinalysis results returned to normal ranges. The serum inorganic fluoride concentration was measured before anesthesia, 1 h after initiation of anesthesia, every 2 h during anesthesia, and again 0, 1, 2, 3, 16, 40, 64, 112, and 160 h after cessation of anesthesia.

Gas samples were obtained for compound A concentration analysis from the inspiratory limbs of the anesthetic circuit distal to the one-way valves via a capped stopcock port, using gas-tight glass syringes (Supelco, Bellefonte, PA). Samples were obtained from the inspiratory limb every 1 h after intubation and at the end of anesthesia. Inspiratory limb gas samples (100 ml) were injected into a sealed evacuated 155-ml sample bottle with stopper and aluminum crimp cap. By injecting 55 ml air, this sealed vacuous sample bottle returned to ambient pressure immediately before the concentrations of compound A in the circuit were measured. Thereafter, gas (200 [micro sign]l) was extracted using a gas-tight syringe and injected into the gas chromatograph (GC-14A, Shimazu, Japan). A glass column with a length of 5 m and an interval diameter of 3 mm packed with 20% dioctyl phthalate on a Chromosorb WAW (GL Science Co., Tokyo, Japan) 80/100 mesh was maintained at 110 [degree sign]C in the gas chromatograph. The injection port was maintained at 130 [degree sign]C. A carrier stream of nitrogen flowing at 30 ml/min was delivered through the column to a hydrogen flame ionization detector. The gas chromatograph was calibrated by preparing standard calibration gases from stock solutions of compound A supplied by Maruishi Pharmaceutical (Osaka, Japan). Briefly, we prepared 1, 10, 40, and 100 ppm compound A by vaporizing stock solutions of compound A in a 155-ml bottle. We extracted 200 [micro sign]l gas from the bottle and injected it into the gas chromatograph using a gas-tight syringe. Calibration curves were linear in the range of 1 - 100 ppm (r²= 0.99).

Urinary NAG activity, [small beta, Greek]^{2-microglobulin}, and clinical laboratory tests, such as BUN, were measured in the clinical laboratories of the Self Defense Force Central Hospital. The clinical laboratories quantitatively assayed the urine for protein and glucose concentrations. These measurements were performed in a single-blinded manner. Serum and urinary inorganic fluoride ions were measured with an ion-selective fluoride electrode and Ionalyzed no. 920 (Orion Research, Boston, MA). Urinary NAG activity (24 h) was determined colorimetrically using a commercially available method (Shionogi, Osaka, Japan). Urinary [small beta, Greek]^{2-microglobulin}concentrations were measured by radioimmunoassay ([small beta, Greek]^2-Micro[middle dot] RIA-BEARS, Dainabot, Tokyo, Japan). Urinary protein or glucose concentration (24 h) was determined by the pyrogallol red or glucose oxidase peroxidase method. The lowest determinable concentration of protein or glucose in our hospital was 5 mg/dl or 50 mg/dl, respectively. Electrophoresis was used to analyze the composition of total protein concentration in urine samples having concentrations >15 mg/dl by autoanalyzer CTE-150N (Jookou, Tokyo, Japan).

The total compound A exposure was calculated from the areas under the curve (AUC) of compound A concentration versus time by using the trapezoid rule (initial concentration was assumed to be zero). [25]Values are presented as mean +/- standard deviation when they were distributed normally. Inter- and intragroup comparisons of laboratory data were performed using two-way repeated measures analysis of variance followed by the Student-Newman-Keuls post hoc test for multiple comparisons. Analyses of patient demographic data and the maximum data, such as serum fluoride concentrations, were performed with one-way analysis of variance followed by the Student-Newman-Keuls post hoc test. Concentrations of urinary excretion of protein and glucose were presented as medians because they were not distributed normally. Comparison of urinary excretion of protein and glucose among the three groups were performed with the Friedman test or Kruskall-Wallis test followed by Dunnett's or Scheffe's F post hoc test. The chi-squared analysis was used to determine whether the appearance rate of proteinuria or glucosuria differed among the three groups. Regression analysis was used to evaluate the correlation between inspired compound A AUC or peak serum fluoride concentration and mean values of several markers after anesthesia using Pearson's product-moment correlation coefficient or Spearman's rank test. Differences were considered significant when P < 0.05.

(Table 1, Table 5) presents the characteristics of the patients in each of the three groups. There were no differences among groups in any measured variable. The individual peak and mean concentrations of compound A in the low-flow sevoflurane group was 41.2 +/- 9.0 ppm and 29.1 +/- 7.1 ppm, respectively. The corresponding value in the high-flow sevoflurane group was 7.2 +/- 4.8 and 3.9 +/- 2.2 ppm, respectively (Table 1). A strong relation was found between compound A concentration and end-tidal sevoflurane concentration in the low-flow sevoflurane group (r²= 0.67; P < 0.001;Figure 1). In contrast, the correlation between compound A concentration and end-tidal sevoflurane concentration in the high-flow sevoflurane group, although significant, was weaker (r²= 0.30; P < 0.001;Figure 1).

Serum fluoride concentrations and urinary excretion of fluoride in both the low-flow sevoflurane and high-flow sevoflurane groups were significantly greater than in the isoflurane group during or after anesthesia at both times (Figure 2). There were no significant differences between the low-flow sevoflurane and high-flow sevoflurane groups with respect to serum fluoride concentrations and urinary fluoride excretion (Figure 2).

(Table 2) lists the normal ranges of several markers among patients in our hospital. Clinical laboratory baseline values did not differ in the three groups, and no abnormal changes in values of renal function studies were noted during the study period; neither elevated BUN and serum creatinine concentrations nor decreased creatinine clearances in any patients was seen (Figure 3, Figure 4, Figure 5).

Results of measurement of 24-h urinary excretion of NAG and [small beta, Greek]^{2-microglobulin}, protein, and glucose concentrations for the three groups before and 1 - 7 days after anesthesia are shown in Figure 6, Figure 7, Figure 8and Table 3, Table 6. Increased NAG excretions were seen in both the low-flow sevoflurane and high-flow sevoflurane groups (Figure 6). There were significant differences between the low-flow isoflurane group and the other two groups with respect to the maximum and mean values for urinary excretion of NAG. However, no significant difference existed in corresponding values between the low-flow sevoflurane and high-flow sevoflurane groups (Table 3). Although urinary excretion of [small beta, Greek]^{2-microglobulin}(24 h) in the low-flow sevoflurane group was significantly greater on days 2 - 5 after anesthesia than before anesthesia, there was no significant difference among the three groups (Figure 7). There were also no significant differences in the maximum and mean values for urinary excretion of [small beta, Greek](^{2-microglobulin}) among the three groups (Table 3). Ten patients in the low-flow sevoflurane group showed 24-h urinary excretion of protein that exceeded the normal range of 150 mg/24 h after anesthesia. In contrast, one patient in the isoflurane group and none in the high-flow sevoflurane group exhibited a urinary excretion of protein >150 mg/24 h (P < 0.01;Table 3). Urinary excretion of protein (24 h) was significantly higher than in both the isoflurane and high-flow sevoflurane group on days 1 - 4 after anesthesia (Figure 8). Furthermore, the maximum 24-h urinary protein excretion in the low-flow sevoflurane group was significantly greater than in the other two groups (P < 0.01;Table 3). Electrophoresis revealed that the excreted protein consisted of approximately 80% albumin, 3%[small alpha, Greek](^1-globulin), 4%[Greek small letter alpha^2-globulin, 6%[small beta, Greek] globulin, and 7%[small gamma, Greek]-globulin. There was no significant difference among the three groups in the number of patients who showed glucosuria or the maximum 24-h urinary excretion of glucose (Table 3), although three patients in the low-flow sevoflurane group showed glucosuria after anesthesia (Table 3; patients 26, 35, and 38); these three patients had the highest urinary excretion of protein in the low-flow sevoflurane group (Table 3). However, the relation between proteinuria and glucosuria evident in the low-flow sevoflurane group was not found between proteinuria and increased eyzymuria or [small beta, Greek]^{2-microglobulinuria}; there was no correlation between urinary excretion of the amount of protein and of the amount of NAG or [small beta, Greek](^{2-microglobulin})(Table 3).

No correlation was found between inspired compound A AUC and creatinine clearance (data not shown). There was also no correlation between peak fluoride concentration and the mean 24-h urinary excretion of [small beta, Greek]^{2-microglobulin}or protein in all patients and between inspired compound A AUC and the mean 24-h urinary excretion of NAG or [small beta, Greek]^{2-microglobulin}in patients who received sevoflurane (Figure 9B, Figure 9C, Figure 9D, Figure 9E). The correlations between peak fluoride concentration and the mean 24-h urinary excretion of NAG in all patients and between inspired compound A AUC and the mean 24-h urinary excretion of protein in patients who received sevoflurane were weak, although significant (r²= 0.20, P = 0.003 versus r²= 0.17, P = 0.003, respectively;Figure 9A, Figure 9F). However, no correlation between peak fluoride concentration and the mean 24-h urinary excretion of NAG and between inspired compound A AUC and the mean 24-h urinary excretion of protein was found in the low-flow sevoflurane group (P = 0.78, P = 0.89).

Values for the serum aspartate aminotransferase and alanine aminotransferase increased slightly in the three groups compared with the preoperative values (Table 4). There were no significant differences in aspartate aminotransferase and alanine aminotransferase values on the postanesthetic days (Table 4) nor in the number of patients who exhibited abnormal value of aspartate aminotransferase or alanine aminotransferase among the three groups. No significant difference existed in the maximum or mean values for aspartate aminotransferase or alanine aminotransferase after anesthesia among the three groups (data not shown).

Urinary excretion of protein increased in patients breathing 1.4 +/- 0.4 minimum alveolar concentration sevoflurane and 29.1 +/- 7.1 ppm compound A for 6.7 +/- 0.9 h. These findings confirm those of Eger et al. [10]in volunteers. We also found that urinary NAG concentration increased after sevoflurane more than after comparable anesthesia with isoflurane. The increases with sevoflurane did not depend on a high concentration of compound A, being equally increased in the low-flow sevoflurane and high-flow sevoflurane groups. These results suggest that inorganic fluoride, which did not differ between the low-flow sevoflurane and high-flow sevoflurane groups, was the likely cause of elevated NAG excretion. These findings are consistent with our previous suggestion that higher concentrations of fluoride are associated with increased urinary excretion of NAG. [26] 

The chemical structure of Compound A contains six fluoride atoms and undergoes cytochrome P-450 - catalyzed defluorination at the fluoromethyl moiety, [9,27]which is similar to the P450-catalyzed defluorination of sevoflurane. Therefore, Eger et al. [10]speculated that defluorination of compound A may increase serum fluoride concentrations. Indeed, urinary excretion of fluoride increases in rats that receive compound A intraperitoneally. [9]However, in the current study no significant difference occurred between the low-flow sevoflurane and high-flow sevoflurane groups with respect to serum and urine fluoride concentrations. These results suggest the in vivo fluoride formation from compound A metabolism is not significant compared with that from sevoflurane.

The site of compound A nephrotoxicity in rats was the renal tubule, especially the tubulus in the region of the outer strip of the outer medullary layer (corticomedullary junction). [5,6,8,9]Compound A nephrotoxicity in rats results in increased urinary excretion of NAG, protein, and glucose. [7–9]Urinary NAG is a lysosomal enzyme that originates from the proximal renal tubules; it is released into the tubular lumen and excreted in the urine after cell necrosis. [28]Proteinuria and glucosuria probably stem from a disturbance of proximal tubular reabsorption, consistent with cell necrosis that has been localized to tubules. [7–9]In contrast to studies of NAG, protein, and glucose, we are unaware of any study that investigated urinary excretion of [small beta, Greek]^{2-microglobulin}in compound A nephrotoxicity in rats. [small beta, Greek]^{2-microglobulin}, a low-molecular-weight protein, is freely filtered through the glomerulus; >99% is reabsorbed by the proximal convoluted tubules, therefore, urinary excretion of [small beta, Greek]^{2-microglobulin}is used as a measure of abnormal tubular function. [28] 

In the current study, increased urinary NAG:creatinine ratios were seen in both the low-flow sevoflurane and high-flow sevoflurane groups. The significant excretion of NAG in the high-flow sevoflurane group corresponds with findings from our previous studies. [26]Twenty-four-hour urinary excretion of protein in the low-flow sevoflurane group was significantly greater than in the other groups. Furthermore, glucosuria was observed in the three patients with the greatest proteinuria, and no patient exhibited glucosuria in both the isoflurane and high-flow sevoflurane groups. These results might suggest that tubular cell damage or tubular dysfunction were present. Furthermore, the result that albumin was 80% of the protein might reflect an event at the level of the glomerulus. However, these changes were not indicative of organ dysfunction, nor were they clinically significant because no patient in the three groups exhibited increased BUN and serum creatinine concentrations and decreased creatinine clearance. This result, that BUN and serum creatinine concentrations were unchanged, is consistent with previous investigations of high-flow sevoflurane [26,29–33]and low-flow or closed-circuit sevoflurane. [10–20] 

The urinary abnormalities in this study were transient, consistent with previous investigations in volunteers [10]and rats. [8]Eger et al. [10]reported that the urinary abnormalities (protein, glucose, and [small alpha, Greek]-glutathione-S-transferase) were greatest 2 or 3 days after anesthesia in volunteers who received low-flow sevoflurane, but most showed a return to normal ranges by 5 - 7 days. Keller et al. [8]reported that compound A-induced nephrotoxicity in rats is reversible; urinalysis changes (protein, glucose, and NAG) and clinical chemistry concentrations (BUN, creatinine) returned to normal by 14 days after exposure, and the severity of pathologic changes that indicated renal necrosis was reduced with time.

Bito et al. [19]measured BUN and serum creatinine concentrations, creatinine clearance, and urinary excretion of NAG on postanesthetic days 1–3 in patients with gastric cancer who underwent gastrectomy. Kharasch et al. [20]measured BUN and serum creatinine concentrations and urinary excretion of NAG, protein, and glucose in patients who underwent primarily lower abdominal surgery. The results of the two studies were the same: there were no differences in values by anesthetic or flow rate. The discrepancy regarding urinalysis changes noted among the results we obtained and those by Eger II et al., [10]Bito et al., [19]and Kharasch et al., [20]might arise from the concentration and exposure time of compound A and the extent of surgical trauma. First, compound A nephrotoxicity in rats is dose dependent, [5,6,18,9]and the dose-dependent effect may be applicable to humans. [11]The inspired compound A AUC in the studies by Bito et al. [19]and Kharasch et al. [20]were 122 ppm/h and 79 ppm/h, respectively, whereas the corresponding values in our study and that of Eger et al. [10]were 192 ppm/h and 328 ppm/h, respectively. Furthermore, the 24-h urinary excretion of protein in the study by Eger et al. [10]was greater than what we observed. Second, the extent of increase in urinary excretion of NAG is proportional to the stress induced by surgery. [34]After minor surgery, urinary NAG excretion does not increase to more than two times the upper limit of normal. [34]The urinary NAG:creatinine ratios after anesthesia for our patients anesthetized with low-flow isoflurane were < 2.3 U/g creatinine, and they did not increase compared with values before anesthesia. Conversely, corresponding values in the study by Bito et al. [19]increased three times, compared with those before anesthesia (2.8 U/g creatinine before anesthesia compared with 10.5 U/g on day 2 after anesthesia). Corresponding values in the study by Kharasch et al. [20]exceeded 5 U/g creatinine. These results indicate that the effect of surgical trauma on urinary excretion of NAG in our patients was less than that experienced by the patients studied by Bito et al. [19]and Kharasch et al. [20]Finally, the substantially higher anesthetic concentrations used in our study and that of Eger et al. [10]may also explain the observed differences in urinalysis changes.

The NAG and protein results must be interpreted carefully, as Kharasch et al. [20]warned. Our results that urinary excretion of protein and NAG increased in the absence of either changes in BUN and serum creatinine concentrations or in creatinine clearances can be interpreted in two ways:(1) tubular or glomerular changes may occur in the absence of organ dysfunction, or (2) the absence of changes of BUN and creatinine concentrations and creatinine clearance may indicate that there was no relevant damage and that the other changes were unimportant. Mazze and Jamison, [35]in an editorial accompanying the articles by Bito et al. [19]and Kharasch et al., [20]also discussed how postoperative renal function should be assessed in patients undergoing surgery. Measurement of BUN and creatinine concentrations are established indices of renal function, reflecting function in the preceding 12 - 24 h, and they are prognostically significant in clinical medicine. [35]In contrast, measurement of urinary enzyme excretion has not been validated adequately as a reliable indicator of clinically significant renal injury in humans. [36]There is also the criticism that urinary enzymes are too sensitive, in that elevations are sometimes present in the absence of other measurable abnormalities. [28]That increased enzymuria denotes tubular cell necrosis is not proved in humans. [36]An increase in NAG concentrations was not seen in combination with an increase in protein concentrations in this study. In contrast to our previous study, [26]in the current study we did not find that patients in whom serum inorganic fluoride concentrations were >50 [micro sign]M had greater urinary excretion of NAG compared with patients in whom serum inorganic fluoride concentrations were <50 [micro sign]M. Consequently, results can be interpreted to support the suggestion that the changes of NAG concentrations were inconsequential. Compared with urinary enzymes, increased urinary excretion of protein is a reliable marker of renal impairment. [35,37]However, proteinuria obviously can occur in completely benign situations and need not be predictive of subsequent renal malady. [37]Albuminuria in the absence of low-molecular-weight proteinuria is considered a specific and sensitive indicator of changes in the determinants of glomerular permeability. [37,38]Indeed, there were no significant differences among the three groups with respect to urinary excretion of [small beta, Greek]^{2-microglobulin}, which is a highly sensitive indicator of proximal tubular function. However, we do not know whether the benign forms of proteinuria apply to the results we found in this investigation.

In conclusion, our study shows that low-flow sevoflurane anesthesia for 6.7 h was associated with increases in urinary protein. In these young, healthy patients without renal disease, this proteinuria was transient and was not associated with changes of BUN concentration, creatinine concentration, or creatinine clearance. However, further studies will be needed to resolve the renal effect of low-flow sevoflurane anesthesia in patients with preexisting renal disease, in those who received potential nephrotoxic drugs, such as aminoglycosides, or in elderly patients.

The authors thank the surgeons and nurses in the operating room and orthopedic wards of our hospitals for their cooperation. They also thank the staff of the clinical laboratories of the Self Defense Force Central Hospital.