Clinical Isoflurane Metabolism by Cytochrome P450 2E1 

Twenty-two patients undergoing anesthesia for elective surgery were studied after providing written informed consent. The investigational protocol was approved by the Institutional Human Subjects Committee. Patient selection criteria and protocol were similar to those used previously. [12-14]Briefly, subjects were within 30% of ideal body weight, had no history of hepatic or renal insufficiency, ethanol abuse, or current use of medications known to alter hepatic drug metabolism, and were undergoing peripheral (i.e., no thoracic or abdominal) surgical procedures of anticipated 8-10 hr duration. Patients were randomized to receive disulfiram 500 mg orally at bedtime on the evening before surgery (N = 12), or nothing (controls; N = 10).

All patients received a standardized anesthetic designed to minimize potential drug interactions other than the desired test interaction. Anesthesia was induced with propofol (1-2.5 mg/kg), fentanyl (50-100 [micro sign]g), and succinylcholine, followed by tracheal intubation, and thereafter maintained with isoflurane (1.5% end-tidal target concentration; 1.3 MAC) for 8 hr, unless the surgical procedure ended before that time. Inspired and end-tidal enflurane concentrations (Capnomac, Datex Medical Instrumentation, Tewksbury, MA) were recorded at 15 min intervals while patients were intubated. For operations lasting longer than 8 hr, isoflurane was discontinued after 8 hr and anesthesia was maintained with propofol, nitrous oxide, and fentanyl. Anesthetic dose was calculated as the product of end-tidal isoflurane concentration (MAC = 1.15%, uncorrected for age) and time, determined in 15 min intervals, which were summed to obtain total isoflurane exposure, expressed as MAC-hr.

Venous samples for determination of blood isoflurane concentration were obtained prior to induction, and at intervals for 24 hr after the start of isoflurane. Samples were frozen at -20 [degree sign]C, and isoflurane concentrations measured by gas chromatography as described. [12,20]Venous samples for determination of plasma trifluoroacetic acid and fluoride concentration was obtained prior to induction, and periodically for 96 hr after the start of isoflurane anesthesia. Samples were frozen at -20 [degree sign]C until analysis. Urine for determination of trifluoroacetic acid and fluoride concentration was obtained for four consecutive 24 hr intervals beginning at the start of anesthesia. Urine was thoroughly mixed, the volume was measured, and an aliquot frozen at -20 [degree sign]C for later analysis. Trifluoroacetic acid concentrations were determined by ion chromatography and fluoride was measured by ion-selective electrode, with limits of quantification of 1 and 2 [micro sign]M, respectively, as described previously. [12,21] 

Patient demographic data and peak plasma fluoride concentrations, were analyzed by Student's unpaired t-test. Urine trifluoroacetic acid and fluoride excretion in the two groups were compared by repeated measures analysis of variance. Data were analyzed using SigmaStat (SPSS Inc., Chicago, IL). Results are expressed as the mean +/− standard deviation.

Patient demographic data are provided in Table 1. Control and disulfiram-treated groups were similar with respect to age, weight, gender, and duration of surgery, as well as blood loss (not shown). The dose of isoflurane delivered, determined from end-tidal isoflurane concentrations, was also similar in both groups. Isoflurane dose during the 8 hr period of anesthetic delivery was 10.2 +/− 1.1 and 9.7 +/− 1.2 MAC-hr in controls and disulfiram-treated patients, respectively, while total dose during the entire anesthetic period was 10.8 +/− 1.1 and 10.1 +/− 1.3 MAC-hr. Blood isoflurane concentrations (approximately 300 [micro sign]M) were not significantly different in controls and disulfiram-treated patients, further confirming equivalency of isoflurane dose.

The pattern and time course of isoflurane metabolites plasma concentration and urinary excretion were consistent with previous reports. [10,22]Average urine fluoride excretion diminished after postoperative day 2, while trifluoroacetic acid elimination was more constant (Figure 1). Plasma trifluoroacetic acid and fluoride concentrations in a representative control patient are shown in Figure 2. Fluoride concentrations were higher than those after 1-3 MAC-hr isoflurane. [22,23] 

Disulfiram pretreatment markedly inhibited isoflurane metabolism, assessed by urine excretion of trifluoroacetic acid and fluoride (Figure 1). In disulfiram treated patients, daily excretion of trifluoroacetic acid and fluoride were significantly diminished on postoperative days 1-4 compared with controls. Trifluoroacetic acid in urine was undetectable on day 1 in 8/12 disulfiram-treated patients, and excretion was 1.5 +/− 2.7 [micro sign]moles, compared with 90 +/− 150 [micro sign]moles in controls. Cumulative 0-96 hr excretion of trifluoroacetic acid and fluoride in disulfiram-treated patients was 34 +/− 72 and 270 +/− 70 [micro sign]moles, respectively, compared to 440 +/− 360 and 1500 +/− 800 [micro sign]moles in controls (P < 0.05 for both).

Disulfiram treatment abolished the rise in plasma trifluoroacetic acid and fluoride concentration following isoflurane administration. Results for representative control and disulfiram patients are shown in Figure 2.

Patient pretreatment with a single disulfiram dose the night before surgery markedly inhibited isoflurane metabolism, as evidenced by substantial reductions in urine trifluoroacetic acid and fluoride excretion, and plasma metabolite concentrations. Cumulative 0-96 hr urinary metabolite excretion was inhibited 82-93%, and plasma metabolite area under the concentration-time curve was diminished more than 90%, compared with controls. Urine collection for additional postoperative days would likely have shown a greater degree of inhibition, due to more complete metabolite recovery, particularly in controls. Disulfiram effects were not attributable to differences in isoflurane dose, which was similar in both groups. Rather, differences between groups were due to disulfiram inhibition of isoflurane metabolism. Disulfiram inhibition of isoflurane metabolism strongly suggests that P450 2E1 is the predominant P450 isoform catalyzing human isoflurane metabolism in vivo.

Single-dose disulfiram has previously been demonstrated to be an effective inhibitor of human P450 2E1 activity in vivo. [18]Oral disulfiram (500 mg) administered 10 hr prior to chlorzoxazone, a specific noninvasive probe of hepatic P450 2E1 activity, [24]significantly diminished P450 2E1 activity in vivo evidenced by an 85% decrease in chlorzoxazone 6-hydroxylation. Disulfiram inhibition of human P450 2E1 in vivo was subsequently used to probe and identify P450 2E1 participation in the metabolism of enflurane, sevoflurane and halothane in humans in vivo, and to diminish the P450 2E1-mediated toxification of halothane. [12-14]Based on urinary metabolite excretion, disulfiram inhibited metabolism of chlorzoxazone 93%, enflurane 89%, halothane 84-91%, isoflurane 82-93%, and sevoflurane 73-80%. Thus P450 2E1 is the predominant isoform which metabolizes halothane, enflurane, isoflurane, and sevoflurane in humans in vivo.

When we previously reported disulfiram reductions of enflurane, sevoflurane and halothane metabolism in humans in vivo, [12-14]the effectiveness of disulfiram inhibition of P450 2E1 activity in vivo had been clearly demonstrated, [18]however, disulfiram specificity toward only P450 2E1 was unknown. Specifically, owing to overlapping substrate specificities of P450 2E1 and 2A6, [25-28]and reports of disulfiram metabolite inhibition of human P450 2A6 in vitro, [25,29,30]it could not be unambiguously established that disulfiram inhibition of anesthetic metabolism represented only P450 2E1-mediated activity, particularly to the exclusion of P450 2A6.

Recent clinical investigations have determined, however, that single-dose disulfiram does exhibit selectivity for P450 2E1 in vivo. Disulfiram did not inhibit human P450 2A6 activity, assessed by the P450 2A6 probe coumarin. [31]Urine 7-hydroxycoumarin excretion was unchanged by single-dose disulfiram, [31]administered using the same protocol as in the present and previous investigations of anesthetic metabolism. Furthermore, additional clinical evaluations have recently shown that disulfiram does not significantly inhibit human P450 2C9, 2C19, 2D6, and 3A4 activities in vivo. [Section] Together with 2E1, these isoforms comprise the majority (approximately 60%) of total human hepatic P450s (of which 73% have been identified). [32]Thus it is apparent that single-dose disulfiram inhibition of drug metabolism is selective toward P450 2E1. Disulfiram inhibition of isoflurane metabolism in the present investigation, and that of enflurane, sevoflurane, and halothane in previous investigations, [12-14]suggests that P450 2E1 is the predominant isoform responsible for metabolism of these volatile anesthetics in humans in vivo.

A characteristic yet previously unexplained feature of single-dose disulfiram inhibition of volatile anesthetic metabolism has been marked inhibition of plasma metabolite appearance and urinary excretion immediately postoperatively, with small increases on postoperative days 3 and 4. [12-14]It is now apparent that this represents resynthesis of P450 2E1 after 48-72 hr. [33]In a recent clinical investigation, P450 2E1 activity was probed by chlorzoxazone clearance and 6-hydroxychlorzoxazone formation clearance before, and 1, 3, 6, 8, 10, and 13 days after disulfiram. [33]P450 2E1 activity returned with a half-life of 32 hr, and was complete after 8 days. Return of activity was attributed to enzyme resynthesis, rather than prolonged elimination of disulfiram or its metabolites. Thus metabolite appearance after 3-4 days in disulfiram-treated patients suggests prolonged low-level anesthetic residence and resynthesis of P450 2E1, as observed with enflurane, halothane, sevoflurane, [12-14]and now isoflurane. If greater inhibition of metabolite formation is desired, a second disulfiram dose should be administered two days postoperatively.

Identification of P450 isoforms responsible for human anesthetic metabolism rests on several approaches. In vitro techniques include (1) metabolism by human cDNA-expressed isoforms, inhibition of microsomal metabolism by (2) isoform-selective antibodies and/or (3) chemical inhibitors, and correlation of drug metabolism with the (4) activity, or (5) protein content of specific isoforms in microsomes from a population of human livers (or other tissues) with varying isoform activities (due to genetic variation, enzyme induction and/or enzyme inhibition). In vivo methods include inhibition of metabolism by drugs which are isoform-selective inhibitors, induction of metabolism, and correlation of metabolism with the hepatic activity or protein content of specific P450s in a population of humans. Our identification of P450 2E1 participation in in vitro human liver microsomal enflurane, methoxyflurane, sevoflurane, and halothane metabolism was based on approaches 1 and 3-5 above, however, only method 3 could be used to implicate P450 2E1 in isoflurane metabolism due to the low rate of isoflurane metabolism. [15,16,20,34]The present clinical investigation corroborates this in vitro implication, and more firmly identifies the role of P450 2E1 in human isoflurane metabolism. Isoflurane metabolism by P450 2E1 explains greater metabolism in patients taking isoniazid, [17]which induces P450 2E1 activity. [35,36]Chronic ethanol consumption also induces P450 2E1 in vivo, [37]increases rat liver microsomal isoflurane metabolism in vitro, [38]and would be expected to enhance isoflurane metabolism in humans. In vitro-in vivo correlations have thus been firmly established for human volatile anesthetic metabolism by P450.

Immune-mediated hepatic necrosis due to isoflurane metabolism and hepatic protein trifluoroacetylation is rare, presumably due to the limited extent of isoflurane metabolism and protein adduction. [2,10]One report, in a patient taking phenytoin, suggested that enzyme induction by phenytoin contributed to isoflurane toxicity, and that patients taking any drug which induces P450 are at increased risk for hepatotoxicity. [5]This attribution and warning are questionable, however, since phenytoin is not known to induce human P450 2E1, the number of P450 2E1 inducers is small, and enzyme induction was not a risk for hepatitis after isoflurane. [3] 

In summary, P450 2E1 appears to be the predominant cytochrome P450 isoform responsible for clinical isoflurane metabolism in humans.

We appreciate the assistance of the various anesthesiologists, anesthesiology residents and nursing personnel who participated in this investigation.

[Section] Single-dose disulfiram had no effect on the activity of P450s 2C9, 2C19, 2D6 or 3A4 in humans, assessed by tolbutamide, mephenytoin and dextromethorphan metabolism, and midazolam clearance, respectively (E.D. Kharasch, manuscript in preparation).