Renal Function in Patients with High Serum Fluoride Concentrations after Prolonged Sevoflurane Anesthesia

Written informed consent for participation was obtained from each patient before study after approval from our hospital ethics committee had been obtained. A total of 34 ASA physical status 1 male patients scheduled to undergo orthopedic surgery expected to last at least 5 h were studied. We selected patients scheduled for peripheral orthopedic surgery without major blood loss, such as knee ligament reconstruction (Table 1and Table 2). A tourniquet was inflated during the operation when the surgical site was in an extremity. Patients with abnormal renal function were excluded; normal renal function was confirmed by routine laboratory renal tests, overnight urine-concentrating test, and determination of urinary NAG excretion. Patients were assigned to receive either sevoflurane (23 patients) or isoflurane (11 patients) anesthesia. 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) to facilitate tracheal intubation. 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. An intraurethral catheter was inserted to facilitate measurement of urinary output. Anesthesia was maintained with sevoflurane or isoflurane, air, and oxygen (FI^O2= 0.3) at a total gas flow of 6 l/min. A semiclosed-circle system with a soda lime canister (Wakolime; Wako Pure Chemical, Osaka, Japan) was used to absorb carbon dioxide. The volatile anesthetic was administered via a Penlon vaporizer (PPV Sigma; Penlon, Abingdon, United Kingdom) or a Muraco vaporizer (Forawic; Muraco Medical, Tokyo, Japan). Anesthesia was maintained for at least 300 min, even if surgery was completed earlier than anticipated. Ventilation was assisted or controlled to maintain carbon dioxide tension at 40 mmHg and arterial oxygen tension greater than 100 mmHg. End-tidal concentrations of sevoflurane or isoflurane were analyzed with a Capnomac Ultima gas analyzer (Capnomac; Datex, Helsinki, Finland), which was calibrated immediately before each study using a cylinder that contained a mixture of gases of known concentrations. The MAC hours for sevoflurane and isoflurane exposures were each calculated from the percent anesthetic concentration and the duration of anesthetic exposure. The MAC values were 2.05% for sevoflurane [7]and 1.15% for isoflurane. [8]Anesthetic concentration was adjusted by the anesthesiologist to maintain systemic arterial blood pressure within plus/minus 20% of baseline.

Serum inorganic fluoride concentration was measured before anesthesia, at 1 h after initiation of anesthesia and then every 2 h during anesthesia, and at 0, 1, 2, 3, 6, 14, 16, 20, 40, and 64 h after cessation of anesthesia. Lactated Ringer's solution was administered 5-6 ml *symbol* kg sup -1 *symbol* h sup -1 during anesthesia and 2 ml *symbol* kg sup -1 *symbol* h sup -1 for 16 h after cessation of anesthetic exposure. Thirty minutes after administration of intravenous fluids had been completed--that is, 16.5 h after the end of anesthetic exposure--each patient received a subcutaneous injection of 10 units of aqueous vasopressin, urine was collected every 30 min for 4 h, and urinary osmolality was determined to evaluate renal-concentrating ability. During vasopressin testing, each patient's oral intake was restricted.

Urine collection began 24 h before anesthesia and continued until 72 h after cessation of anesthesia. Before operation and on days 2 and 3 after anesthesia, overnight urine-concentrating ability was determined by restricting oral intake beginning at 8 PM and obtaining one urine specimen at 6 AM the next morning and another 1 h later. The osmolality of the 7 AM specimen was determined. Clinical laboratory studies were performed immediately before anesthesia and repeated 24, 48, and 72 h after initiation of anesthesia.

All patients received antibiotics perioperatively. The specific antibiotic was determined by the patient's surgeon, who was not involved in the study. No patient received aminoglycosides (33 patients received cephalosporins and one patient aspoxicillin; Table 2). Antibiotics were administered intravenously twice a day from immediately after the induction of anesthesia to day 2 or 3 after anesthesia, and thereafter 600 mg cefotiam was administered orally for 3-5 days. Only two patients (patients 17 and 22; Table 2) received cefotitam prophylactically (600 mg) for 4 or 5 days before operation because of a puncture wound of the knee.

Serum inorganic fluoride ion was measured with an ion-selective fluoride electrode and Ionalyzer No. 901 (Orion Research, Boston, MA). Serum and urinary osmolality were measured with a Model 3D3 osmometer (Advanced Instruments, Norwood, MA) using a freezing point depression test. Urinary NAG activity was determined by spectrophotometric assay using sodio m-cresolsulfonphthaleinyl N-acetyl-beta-D-glucosaminide as described by Noto et al. [9]Urine enzyme activity was expressed relative to creatinine concentration as units of NAG activity/g *symbol* creatinine. In our hospital, the normal range of this parameter is 0.04-2.85 U/g *symbol* creatinine.

Values are presented as mean plus/minus SE. Subjects receiving sevoflurane were divided into two groups based on whether peak inorganic fluoride concentration exceeded 50 micro Meter (sevoflurane^highgroup) or was less than 50 micro Meter (sevoflurane^lowgroup). Differences between the three study groups were analyzed with ANOVA using Scheffe's F procedure. P values less than 0.05 were considered to indicate statistical significance. A power analysis was performed to determine the possibility of type II error with the JMP statistical software package (SAS Institute, Inc., Cary, NC) for Macintosh computers.

Eight patients receiving sevoflurane had a peak fluoride concentration greater then 50 micro Meter. Therefore, there were 8 patients in the sevoflurane^highgroup and 15 in the sevoflurane^lowgroup. Table 1and Table 2list the clinical characteristics of the patients studied. There were no hypotensive episodes in any of the patients. The three groups of patients were similar in clinical characteristics, with the exception of MAC hours. Mean MAC hours in the sevoflurane^highgroup were significantly greater than in each of the other two groups (P < 0.01, sevoflurane^highvs. isoflurane; P < 0.05, sevoflurane^highvs. sevoflurane^low).

Urinary NAG excretion can increase in a variety of circumstances, including after the administration of various types of antibiotics [10]and surgery. [11]Therefore, we analyzed the data separately for 25 patients who received cephalosporins and underwent tourniquet inflation (7 for isoflurane, 7 for sevoflurane^high, and 11 for sevoflurane^low) to reduce the variety of circumstances that might influence urinary NAG excretion. Neither the duration of tourniquet inflation nor the daily dose of cephalosporins administered differed significantly among the three groups. Among these 25 patients, there was no significant difference in mean MAC hours among the three groups (Table 3). The mean peak serum fluoride concentration in the sevoflurane^highgroup was 55.8 plus/minus 3.4 micro mol/l, observed 1 h after cessation of anesthesia (Figure 1). The mean time that peak fluoride concentrations exceeded 50 micro mol/l in this group was 4.6 plus/minus 1.6 h. The mean peak serum fluoride concentration in the sevoflurane^lowgroup was 36.8 plus/minus 1.9 micro mol/l, observed at cessation of anesthesia (Figure 1). The corresponding value in the isoflurane group (n = 11) was 4.8 plus/minus 0.5 micro mol/l, observed 16 h after cessation of anesthesia (Figure 1).

The results of clinical laboratory studies for the three groups are shown in Table 4. The three groups did not differ in laboratory baseline values, and no abnormal changes in values of renal function studies were noted during the study period.

Neither urinary osmolality before vasopressin administration nor the total amounts of fluids administered during and after anesthesia differed among the three groups (Table 1and Table 2). The mean maximum urinary osmolalities after injection of vasopressin in the isoflurane, sevoflurane^high, and sevoflurane^lowgroup were 816 plus/minus 37, 681 plus/minus 60, and 811 plus/minus 32 mOsm/kg, respectively (Table 2, Figure 2). Although mean maximum urinary osmolality in the sevoflurane^highgroup tended to be lower than that in each of the other two groups, the overall difference marginally failed on one-way ANOVA to reach statistical significance (P = 0.068, statistical power = 53%). Power analysis indicated that at least four additional patients were required to obtain a significant difference with the same standard errors and structural results as the current sample. The two patients who exhibited the lowest and second-lowest maximum urinary osmolality belonged to the sevoflurane^highgroup (patients 12 and 17; Table 2, Figure 2). The lowest maximal osmolality in the group of all patients was 390 mOsm/kg, in a patient given sevoflurane whose peak serum fluoride concentration was 56.8 micro mol/l and whose fluoride concentration remained greater than 50 micro mol/l for 6 h (patient 17). There was a significant, albeit weak, inverse correlation between peak fluoride concentration and maximal urinary osmolality after the injection of vasopressin in patients who received sevoflurane (r = -0.42, P < 0.05).

All patients were able to concentrate urine effectively after overnight dehydration (Table 2). In the three groups, urinary osmolality did not differ on day 2 or 3 after anesthesia from the value obtained before anesthesia. In addition, there were no significant differences in overnight urine-concentrating ability among the three groups. Urinary osmolalities after overnight dehydration of the one patient who exhibited impairment of renal-concentrating ability (patient 17) was 815 mOsm/kg before anesthesia, 846 mOsm/kg on day 2 after anesthesia, and 963 mOsm/kg on day 3 after anesthesia (Table 2).

The results of measurement of urinary excretion of NAG for the three groups before and 1, 2, and 3 days after anesthesia are shown in Figure 3. In the isoflurane group, there was no difference between urinary excretion of NAG during the 3-day period after anesthesia and that before anesthesia. Urinary excretion of NAG in the sevoflurane^highand sevoflurane^lowgroups was significantly greater on days 2 and 3 after anesthesia than before anesthesia, and was also significantly greater than that of the isoflurane group on day 2 after anesthesia. The maximum urinary excretion of NAG during the 3-day postoperative period for sevoflurane^highwas significantly greater than that of the isoflurane group (P < 0.05; Figure 4). This significant difference was recognized even in patients limited to those who were administered cephalosporins and underwent tourniquet inflation (P < 0.05; Figure 4). The patient whose serum fluoride concentration was highest (86.8 micro mol/l) had the highest maximum urinary excretion of NAG (20.08 U/g *symbol* creatinine) of all patients studied (patient 13; Table 1). No correlation was found between maximum urinary osmolality after injection of vasopressin and maximal urinary excretion of NAG. Urinary NAG excretion by four patients (patients 16, 23, 25, and 33; Table 2) was measured until day 7 after anesthesia. In these cases, urinary NAG excretion after anesthesia peaked by postanesthesia day 3, and urinary NAG subsequently returned to normal levels by day 6 after anesthesia (values for these four patients returned to normal levels by days 4, 6, 4, and 3 after anesthesia, respectively).

The renal effect of anesthetics would be best studied in volunteers not subjected to surgery. [12]However, we believe the effects of surgical trauma and hemorrhage on renal function in the current study were minimal and were probably the same in the three groups, because the sites of surgery in our healthy patients were almost always in the extremities, and blood loss was negligible. In addition, we made an effort to maintain stable hemodynamics. Because we used a Penlon vaporizer, which permits a maximum anesthetic concentration of 7%, we were able to administer higher concentrations of sevoflurane in the current study than in our previous study. We were, therefore, able to obtain patients whose peak inorganic fluoride concentration exceeded 50 micro mol/l.

The total dosage of anesthetic in the sevoflurane^highgroup was significantly greater than that in the other two groups (Table 1). However, when only those patients who underwent tourniquet inflation were considered, no significant difference was found among the three groups in total anesthetic dosage (Table 3). Although MAC hours in the sevoflurane^highgroup tended to be greater than in the other two groups, it is unlikely that poor renal perfusion caused by greater concentration of anesthetics occurred in the sevoflurane^highgroup, because we adjusted individual anesthetic concentrations to maintain stable hemodynamics.

Soda lime, which was used in this study, converts sevoflurane to an olefin referred to as compound A, [13,14]which is nephrotoxic in rats. [15,16]The threshold of compound A for renal tubular necrosis in rats is 25-50 ppm [15,16]and, thus, is within the range of concentrations that may be found in clinical practice. [17,18]However, the possibility of breathing breakdown products by soda lime is minimized when the fresh gas flow is maintained as high as 6 l/min in a semiclosed circuit, the conditions we employed in the current study. [14]Frink et al. [19]reported that 9.5 MAC hours sevoflurane anesthesia in a semiclosed circuit at a flow rate of 5 l/min resulted in mean peak concentrations of compound A of 7.6 plus/minus 1.0 ppm. Although we did not measure the concentration of compound A in this study, it was presumably less than 25-50 ppm, the level of potential toxicity in rats. It therefore seems unlikely that compound A could have contributed to production of the transient abnormalities in urinary concentrating ability and NAG excretion detected in this study.

Mazze et al. [12]and Frink et al. [5]compared responses to vasopressin or desmopressin before and after prolonged anesthesia in volunteers. We were unable to perform vasopressin tests before anesthesia because of limitations created by the length of the preoperative period of study. Overnight preoperative and postoperative urine-concentrating ability testing was, therefore, substituted for comparison of preoperative with postoperative urine-concentrating ability. Aqueous vasopressin is now widely used in the differential diagnosis of polyuria, because it has a duration of action of 2-6 h. [20]However, the principal use of desmopressin is in treatment of diabetes insipidus, because it has prolonged antidiuretic effects lasting 6-24 h. [20]It is usually administered intranasally, a route featuring greater variability in absorption than intramuscular administration. We therefore believe that aqueous vasopressin may be better suited to early detection of decrements in renal-concentrating ability on day 1 after anesthesia than is desmopressin. For the control group, we chose exposure to isoflurane, which undergoes an insignificant degree of biotransformation to inorganic fluoride.

Although mean maximum urinary osmolality tended to be lower in the sevoflurane^highgroup than in other groups, power analysis revealed that our inability to detect a difference among the three groups probably resulted from a type II error. Patient 17, who had the lowest maximum urinary osmolality, appeared to have abnormal renal-concentrating ability, because his value was more than 3 SD less than the mean of the control isoflurane group. This patient's renal function returned to normal by 2 days after anesthesia, because urinary osmolality after overnight dehydration on day 2 after cessation of anesthesia was greater than that measured preoperatively.

Although the extent of nephrotoxicity of methoxyflurane has been found to be correlated with its dosage and peak serum fluoride concentrations, susceptibility to nephrotoxicity varies in individual patients who receive the same methoxyflurane dosage. [1]Drug interaction, genetic heterogeneity, preexistence of renal disease, and a host of other nephrotoxicity factors may account for the different susceptibility to nephrotoxicity observed among patients. [21]Patient 17 had suffered a sports-related injury, and his knee was swollen and the joint was aspirated because of hemarthrosis. Although he had no signs of infection, he prophylactically received 600 mg cefotitam orally each day for 4 days before anesthesia. The results of preoperative renal function testing of patient 17 were normal, as confirmed by laboratory renal tests, an overnight urine-concentrating test, and determination of urinary excretion of NAG (Table 1). Preoperative administration of antibiotics may have contributed to the difference in concentrating ability between patient 17 and the two other patients who had similar or greater peak fluoride concentrations (patients 13 and 16). Cephalosporins do not commonly have nephrotoxic effects at therapeutic doses, although they are potentially nephrotoxic. [10,22,23]Furthermore, patient 17 received only a low dose of cefotitam. Consequently, we believe it unlikely that preoperative administration of cefotitam contributed to the concentrating defect in patient 17, although we cannot exclude this possibility entirely.

In neither the study by Frink et al. [5]nor our own previous study [6]did any patient anesthetized with sevoflurane exhibit an abnormality in renal-concentrating ability. This may have been the case because of the difference in antidiuretic activity between desmopressin and aqueous vasopressin, and also because only one patient in our study and three patients in the study by Frink et al. [5]had a peak serum inorganic fluoride concentration greater than 50 micro mol/l. Mazze et al. [12]found that a decrease in maximum urinary osmolality occurred after injection of vasopressin tannate in every subject who had undergone prolonged anesthesia with enflurane, although the mean peak fluoride concentration was 33.6 micro mol/l. In the study by Frink et al., prolonged enflurane anesthesia (9.5 MAC hours) produced a renal-concentrating deficit in two of seven subjects with mean fluoride concentrations of 26 micro mol/l. The shape of the time-concentration curve for serum fluoride cannot, by itself, explain why prolonged enflurane anesthesia results in a renal-concentrating deficit despite fluoride concentrations less than those induced by sevoflurane anesthesia. Other factors, such as the effect on renal perfusion of anesthetics proposed by Frink et al., [5]may explain this finding.

Urinary enzymes are much more sensitive indicators of antibiotic-induced nephrotoxicity than are endogenous creatinine clearance, [24]urinary osmolality [25]in rats, or urinary specific gravity in humans. Urinary N-acetyl-beta-D-glucosaminidase (NAG), a lysosomal enzyme originating from the proximal renal tubules, is a sensitive and noninvasive indicator of renal tubular damage. [10,11]Increased urinary excretion of NAG is observed in various renal diseases, in drug-induced renal damage, after surgery, and during episodes of rejection after renal transplantation. [10,11]Urinary excretion of NAG correlated most closely among urinary enzymes with the dose of antibiotic used. [10]Correlations were found between the degree of excretion of NAG and impairment of concentrating ability in dogs [26]and rats [27]with papillary necrosis induced by ethyleneimine. Thus, urinary excretion of NAG is a sensitive indicator of drug-induced renal abnormalities. [10]The extent of increase in urinary excretion of NAG is proportional to the stress induced by surgery. [11]After minor surgery, urinary NAG activity does not increase to more than twice the upper limit of normal. [11]Indeed, the urinary NAG/creatinine ratios for our patients anesthetized with isoflurane did not exceed twice the upper limit of normal. All of our patients received antibiotics. Cephalosporins, as well as aminoglycosides, have been reported to increase urinary NAG excretion. [28,29]In addition, we do not know whether limb ischemia was a factor responsible for the increase in urinary excretion of NAG. Therefore, we compared the urinary excretion of NAG only in those patients who received cephalosporins and underwent tourniquet inflation. Because the maximum urinary NAG/creatinine ratio in the sevoflurane^highgroup was significantly greater than that in the isoflurane group, even for this limited set of patients, it is unlikely that limb ischemia, type of surgery, or antibiotic administration accounted for the increase in maximum urinary excretion of NAG in the sevoflurane^highgroup in this study.

It is of note that it was not on day 1, but rather on days 2 and 3 postanesthesia, at which time the serum fluoride ion concentrations had returned to normal, that significant elevation of urinary NAG excretion occurred. This finding agrees with those obtained by Motuz et al. [30]These authors evaluated the effect of enflurane in surgical patients on urinary excretion of alanine aminopeptidase (AAP), an enzyme found in the brush border membrane of the proximal renal tubule. Urinary AAP excretion in patients anesthetized with enflurane significantly increased to greater than preoperative values not 24 but 48 h after surgery. [30]Although differences in the kind of urinary enzymes tested and anesthetics used between our study and theirs exist, increased urinary excretion of renal tubular enzymes occurred on day 2 postanesthesia. This type of delay in the increase of urinary enzymes was reported by Shimada et al., [31]who found that urinary NAG did not increase until 12 h, but was increased in 12-24-h urine specimens and reached a maximum value within 48 h after injection of inorganic mercury in rats. They suggested that the release of NAG into urine or significant lysosomal degradation occurred at a later time period than mercury damaged proximal tubules. These findings indicate that urinary enzymes in renal tubules do not increase immediately after the serum concentration of the toxin responsible for renal damage reaches a peak value, and that there is a delay in the increase of excretion of urinary enzymes in renal tubules.

In the study by Frink et al., [5]the urinary NAG/creatinine ratio did not increase after prolonged sevoflurane anesthesia. We speculate that this discrepancy between results is caused by the difference in the methods used for urine collection and the use of antibiotics. Frink et al. did not mention the duration of urine collection for measurement of urinary NAG excretion. Notably, 24-h samples are best for precise evaluation of urinary NAG excretion. [10]In the current study, urine was collected continuously for 72 h after anesthesia, and 24-h samples were used for evaluating urinary NAG excretion. However, Frink et al. performed a 24-h urine collection only on day 4 after anesthesia for measurement of creatinine clearance and not of urinary NAG/creatinine ratios. Our patients received antibiotics until at least 3 days after anesthesia. Cephalosporin therapy has been found to increase urinary excretion of NAG. [28]In the current study, the isoflurane group, which received cephalosporins to an extent similar to the sevoflurane group, had no increase in NAG excretion during the 3-day postanesthesia period, indicating that that increases in urinary NAG excretion were caused by sevoflurane anesthesia. It is possible, however, that cephalosporins potentiated fluoride-induced renal damage, and that, as a result, urinary NAG excretion in both the sevoflurane^highand sevoflurane^lowgroups significantly increased after anesthesia.

Our study demonstrates that sevoflurane administration was associated with a dose-related increase of urinary NAG excretion and a transient, significant defect in concentrating ability in one patient and the tendency toward a transient concentrating defect in a group of patients exposed to a high dose of sevoflurane. In these young, healthy patients without renal disease, the results were inconsequential. However, further studies will be required to establish the safety of sevoflurane anesthesia in patients with preexisting renal disease.

The authors thank the surgeons and nurses in the operating room and orthopedic wards of their hospitals, for their cooperation; Professor Kono, Department of Public Health, National Defense Medical College, for his assistance with statistical analysis; and R. Endo, for his assistance in measuring urinary N-acetyl-beta-glucosaminidase excretion.