SINCE the introduction of halothane into clinical practice, fluorinated volatile anesthetics have been associated with liver injury. Although the mechanism of the liver injury caused by fluorinated volatile anesthetics has not been established, evidence suggests it may be due to the oxidative metabolism of these drugs by hepatic cytochrome P450s (CYP) to form immunogenic acylated hepatic protein conjugates. 1In susceptible patients, these modified proteins are believed to induce the formation of pathogenic antibodies and T cells directed against antigenic determinants of these proteins in either their native (autoantigens) or acyl-modified states (neoantigens). We report a case of fulminant hepatic failure and death in association with isoflurane anesthesia 22 yr after previous exposure to enflurane. High concentrations of CYP2A6 autoantibodies were detected in the serum of the patient.
Case Report
A 55-yr-old obese woman was admitted to the hospital after experiencing right upper quadrant pain. The patient’s general health had been good, and there was no history of liver or gastrointestinal problems. She had never received a blood transfusion, she denied intravenous drug use, and she had no history of alcohol use and no recent history of travel or exposure to hepatotoxins. Her only medications were estrogen and progesterone supplements. She had no known allergies. Twenty-two years previously, she had a tubal ligation during general anesthesia with enflurane, which was complicated by persistent postoperative nausea and vomiting.
Abdominal ultrasonography showed cholelithiasis. Endoscopic retrograde cholecystopancreatography was consistent with recent passage of a stone through the ampulla of Vater. Serum biochemical analysis revealed the following: lipase, 1,271 (reference, 56–239) U/l; amylase 252 (35–115) U/l; total bilirubin, 0.7 (0.1–1.1) mg/dl; direct bilirubin, 0.2 (0–0.3) mg/dl; alkaline phosphatase, 206 (90–234) U/l; increased aspartate transaminase, 84 (12–31) U/l; and alanine aminotransferase concentration, 74 (9–29) U/l. The patient’s international normalized ratio was 1.0. A diagnosis of acute biliary pancreatitis was made, and the patient underwent laparoscopic cholecystectomy during general anesthesia. The anesthetic regimen included intubation with sodium thiopental and succinylcholine followed by isoflurane in air–oxygen, fentanyl citrate, cisatracurium, and ondansetron. Anesthesia continued without incident. Specifically, there were no episodes of perioperative hypoxemia or hypotension. At surgery, the biliary anatomy was normal, and evidence of moderate inflammation of the gallbladder was noted. The patient’s immediate postoperative course was remarkable only for abdominal pain and nausea. Liver function tests on the day after surgery showed a total bilirubin concentration of 0.8 mg/dl and an aspartate transaminase concentration of 108 U/l. Discharge from the hospital was delayed by 1 day because of nausea, but the patient went home 3 days after surgery.
One day after discharge, the patient returned to the hospital, reporting nausea, vomiting, abdominal and back pain, tachypnea, and diaphoresis. Laboratory examination revealed the following: total bilirubin, 14.3 mg/dl; direct bilirubin, 8.0 mg/dl; aspartate transaminase, 14,493 U/l; alanine aminotransferase, 6,564 U/l; alkaline phosphatase, 691 U/l; and international normalized ratio, 5.6. Her blood pH was 7.09, with a base deficit of −22 mEq/l. Her lactate concentration was 18.6 (reference, 0.93–1.65) mm, and amylase and lipase concentrations were only mildly increased at 750 and 612 U/l, respectively. The abdominal computed tomography scan and abdominal radiograph were normal. Hepatic ultrasonography revealed normal hepatic vasculature. A diagnosis of fulminant hepatic failure was made, and the patient was admitted to the intensive care unit where, shortly thereafter, stage three hepatic encephalopathy developed. Her trachea was intubated, mechanical ventilation was initiated, and an intracranial pressure monitor was placed. The option of liver transplantation was discussed with the family and was declined.
Investigations for a cause of the fulminant hepatic failure included the following: normal or negative serum ceruloplasmin; anti–smooth muscle, antimitochondrial, and antinuclear antibodies; α1 antitrypsin, anti–hepatitis A virus immunoglobulin (Ig) M, anti–hepatitis C virus, human immunodeficiency virus, and human T-cell lymphotrophic virus screening; and cytomegalovirus IgM. Her anti–Epstein-Barr nuclear antigen was positive, but her anti–viral capsid antigen–IgM was negative. The IgG and IgM varicella zoster titers were low. Her anti–hepatitis B surface antigen was positive, but the rest of her B serology was negative. Her acetaminophen concentration was 18 (reference, < 50) μg/ml. The patient had taken Extra Strength Tylenol tablets (McNeil Consumer Products Co., Ft. Washington, PA) on the day of discharge from the hospital and on the morning of readmission. Her family did not believe that she exceeded the recommended dosage, and the acetaminophen concentration was determined from a sample taken less than 12 h after the last dose. Serum iron concentration was increased at 203 (35–145) μg/dl, and she had a low total iron binding capacity of 208 (250–400) μg/dl. Continuous hemofiltration for renal failure was started. Ten days after cholecystectomy, the patient’s family requested discontinuation of all support, and the patient died. Permission for an autopsy was denied by the family, and neither an ante nor a post mortem liver biopsy was performed.
Serum was collected on hospital day 8 and tested for reactions of serum antibodies to liver microsomes from halothane-exposed and control rats, purified human CYP2E1, 2and human CYP2A6 3by previously described enzyme-linked immunosorbent assays. 2,4Control patients were normal subjects (n = 10) without a history of anesthetic exposure. The enzyme-linked immunosorbent assays revealed that the patient had higher concentrations of serum antibodies that reacted with both trifluoroacetic acid (TFA)–labeled and unmodified rat liver microsomal proteins and CYP2A6 than those of control patients (table 1). In contrast, very little immune reactivity was seen against CYP2E1.
Table 1. ELISA Determination of Serum Antibodies
Enzyme-linked immunosorbent assay (ELISA) screening of patient and control sera for antibodies that react with rat liver microsomal protein from halothane-exposed and control rats (100 μg), purified native human CYP2E1 (1 μg), and purified native human CYP2A6 (1 μg). Sera were diluted 1/100. A reaction was considered positive when the absorbance at 405 nm was greater than 2 SDs above the mean value obtained from the sera of 10 control patients.
* Significantly different from control, P < 0.01, t test.
CYP = cytochrome P-450.
Discussion
Soon after the introduction of halothane, several cases of hepatotoxicity were linked to this agent. 5The hepatic injury may be of two types; a relatively common, mild dysfunction or a rare, severe centrilobular necrosis, which can lead to fulminant hepatic failure. Oxidative metabolism of halothane in the liver, particularly by CYP2E1, produces trifluoroacetyl chloride. This reactive metabolite binds covalently with liver microsomal proteins to form several TFA-labeled neoantigens. 2,6It is believed that these neoantigens may be immunogenic in susceptible patients and induce the formation of specific pathogenic antibodies and T cells directed against native or acyl-modified epitopes of these proteins. 2,7–9Other volatile fluorinated agents are also metabolized to form acylated protein adducts, but at much slower rates than halothane (halothane >>> enflurane > isoflurane > desflurane). 1The incidence of hepatic injury caused by these drugs follows the same trend. 10–15
In a previous study, 67% of a group of 24 halothane hepatitis patients had specific serum reactivity against trifluoroacetylated liver microsomal proteins from halothane-treated rats. 4That number increased to 79% when specific TFA-labeled proteins purified from rat liver microsomes were used as test antigen in the enzyme-linked immunosorbent assay. Importantly, a significant number of halothane hepatitis patients did not react with the trifluoroacetylated protein neoantigens. This observation led us to discover that in many cases, halothane hepatitis was associated with serum antibodies directed against native as opposed to trifluoroacetylated proteins. These proteins include protein disulfide isomerase, 16P 58, 17ERp72, 18and CYP2E1. 2
The P450 isoenzymes are a superfamily of hemoproteins that catalyze the metabolism of a large number of endogenous and exogenous compounds. 19Autoantibodies against specific human CYPs have been found in the sera of patients with a variety of diseases, including those caused by drugs. In the cases of hepatitis caused by tienilic acid and dihydralazine, patients had serum antibodies directed against CYP2C9 and 1A2, respectively. 20,21In a subset of patients with idiopathic autoimmune chronic active hepatitis, antibodies to CYP2D6 were identified. 22Antibodies against CYP1A2 and 2A6 have been found in the sera of patients with autoimmune polyglandular syndrome type I. 23Autoantibodies to CYP2E1 have been previously reported in 45–70% of halothane hepatitis patients’ sera and are thought to be induced by the formation of immunogenic TFA-CYP2E1 adducts. 2,24Cytochrome 2A6 can also oxidatively metabolize halothane 25to form trifluoroacetylated protein adducts, including TFA-CYP2A6. 26In susceptible patients, it is possible that TFA-CYP2A6 may bypass the immunologic tolerance that normally exists against CYP2A6 and may induce the formation of CYP2A6 autoantibodies. 2
Immunologic cross-sensitization has been associated with the use of fluorinated volatile anesthetics. 27Therefore, exposure to one fluorinated volatile anesthetic may sensitize a patient to subsequent exposure by another fluorinated volatile anesthetic and lead to a severe immunologic reaction. Examples of possible cross-sensitization reactions between halothane and enflurane, 28isoflurane, 29or desflurane 15have been previously reported.
This patient’s clinical presentation in many ways was typical of that of patients who have postoperative hepatic injury after exposure to fluorinated inhaled anesthetics. She was a middle-aged, obese woman who presented within a few days of surgery with nausea, vomiting, jaundice, and an unexplained hepatic injury after isoflurane anesthesia. Rash and eosinophilia were not seen. In addition, the search for alternative causes of fulminant hepatic failure ruled out other major precipitating causes. Although hepatitis is known to occur in biliary surgery and pancreatitis, based on this patient’s history, laboratory data, and clinical presentation, we believe instead that she may have been sensitized to enflurane 22 yr previously and that the subsequent isoflurane anesthesia resulted in a cross-sensitization immunologic reaction leading to hepatic failure and death. In a previous report, a patient died after a period of 28 yr between halothane exposures. 30These findings suggest that there may be no safe interval between exposures to fluorinated volatile anesthetics in sensitized patients. Although CYP2A6 antibodies have been identified in association with other immune syndromes, we believe this is the first report of autoantibodies to CYP2A6 in association with drug-induced hepatic injury. The diagnostic and prognostic significance of the CYP2A6 autoantibodies remains to be determined.
The authors thank Allan Rettie, Ph.D., William F. Trager, Ph.D., and Sidney D. Nelson, Ph.D., of the Department of Medicinal Chemistry, School of Pharmacy, University of Washington, Seattle, Washington, for providing the human purified CYP2E1 and CYP2A6.
