Anesthesia & Analgesia

© 2002 by International Anesthesia Research Society.
Volume 94(2)             February 2002             pp 243-249
Autoantibodies Associated with Volatile Anesthetic Hepatitis Found in the Sera of a Large Cohort of Pediatric Anesthesiologists
[Pediatric Anesthesia]

Njoku, Dolores B. MD*‡,; Greenberg, Robert S. MD*†,; Bourdi, Mohammed PhD‡,; Borkowf, Craig B. PhD§,; Dake, Elizabeth M. MS*,; Martin, Jackie L. MD*‡,; Pohl, Lance R. Pharm D, PhD‡

Departments of *Anesthesiology and Critical Care Medicine and †Pediatrics, The Johns Hopkins Medical Institutions, Baltimore, Maryland, the ‡Molecular and Cellular Toxicology Section of the Laboratory of Molecular Immunology, and the §Division of Epidemiology and Clinical Applications, Office of Biostatistics Research, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland
DBN is supported, in part, by the Foundation for Anesthesiology Education and Research.
Presented, in part, at the Fourth Annual Society of Pediatric Anesthesiology Winter Meeting, February 14, 1998, Phoenix, AZ and at the International Anesthesiology Research Society 73rd Clinical and Scientific Congress, March 13, 1999, Los Angeles, CA.
October 9, 2001.
Address correspondence and reprint requests to Dolores B. Njoku, MD, Department of Anesthesiology and Critical Care Medicine, The Johns Hopkins Hospital, Blalock 906A, 600 N Wolfe St., Baltimore, MD 21287. Address e-mail to dnjoku@jhmi.edu.

Outline


Graphics

Abstract^

Anesthetic-induced hepatitis is thought to have an immune-mediated basis, in part because many patients who develop hepatitis have serum autoantibodies that react with specific hepatic proteins. The present study shows that pediatric anesthesiologists also have these serum autoantibodies. Moreover, levels of these autoantibodies are higher than those of general anesthesiologists. We collected sera from 105 pediatric and 53 general anesthesiologists (including 3 nurse anesthetists), 20 halothane hepatitis patients, and 20 control individuals who were never exposed to inhaled anesthetics. Serum cytochrome P450 2E1 (P450 2E1) and 58-kd hepatic endoplasmic reticulum protein (ERp58) autoantibodies were measured by enzyme-linked immunosorbent assays. Positive values were 2 sd above median control values. Two multiple regression models were constructed. Pediatric anesthesiologists, like halothane hepatitis patients, had higher serum autoantibody levels of ERp58 and P450 2E1 than general anesthesiologists and controls, which was possibly because of their increased occupational exposures to anesthetics. Female anesthesiologists had higher levels of ERp58 autoantibodies than male anesthesiologists, whereas female pediatric anesthesiologists had higher levels of P450 2E1 autoantibodies than all other anesthesiologists. One female pediatric anesthesiologist had symptoms of hepatic injury. Because most anesthesiologists do not develop volatile anesthetic-induced hepatic injury, the findings suggest that pathogenic ERp58 and P450 2E1 autoantibodies may not directly cause volatile anesthetic hepatitis. Female anesthesiologists have high levels of these autoantibodies; however, the majority of these individuals do not develop hepatitis, suggesting that autoantibodies may not have a pathological role in volatile anesthetic-induced hepatitis.


Hepatic injury after environmental exposure to volatile, halogenated anesthetic gases is a rare event. However, it has been described in operating room personnel, recovery room nurses, and in laboratory workers anesthetizing laboratory animals (1–4). With regards to operating room personnel, most events have been described before the introduction and widespread use of scavenging systems (5). Despite this precaution, pediatric anesthesiologists can still be exposed to significant levels of these anesthetics during frequent mask inductions of general anesthesia and from the practice of using uncuffed endotracheal tubes.

Previous investigations have demonstrated that several specific hepatic proteins become covalently trifluoroacetylated (TFA) by the reactive metabolites of halothane and isoflurane (6). Similar adducts may also be formed at very small levels after exposure to desflurane (6). These TFA neoantigens are important because it is thought that they induce immune responses against either the TFA neoantigens, the native protein components (autoantigens) of the TFA neoantigens, or both of these classes of antigens in individuals that are susceptible to inhaled anesthetic-induced hepatitis (7). For example, studies in rats have shown that after exposure to halothane, a 58-kd hepatic endoplasmic reticulum protein (ERp58) becomes covalently modified by the trifluoroacetyl chloride metabolite of halothane (8). Subsequently, it was found that the majority of halothane hepatitis patients had serum antibodies that reacted with the purified rat TFA-ERp58 neoantigen, native ERp58 autoantigen, or both antigens, and this reactivity was significantly more than that of control patients (9). Moreover, we recently found that 40% of halothane hepatitis patients have serum autoantibodies that react with human liver ERp58 (Martin et al., unpublished results). Similarly, it has been established that cytochrome P450 2E1 (P450 2E1), the primary enzyme responsible for the oxidative metabolism of most volatile anesthetics, also becomes TFA altered when it metabolizes halothane (10). In addition, autoantibodies reacting with P450 2E1 are significantly elevated in the sera of 45%–70% of patients diagnosed with halothane hepatitis, whereas control subjects did not demonstrate increased levels of these autoantibodies (10,11). These findings suggest that pathogenic antibodies directed against ERp58, P450 2E1, or both may have a role in the etiology of volatile anesthetic hepatitis.

Low levels of P450 2E1 autoantibodies have also been found in the sera of a small group of anesthesiology personnel, suggesting that they may have been formed as a result of environmental exposure to halothane, enflurane, isoflurane, or possibly desflurane (10). If this were the case, then it seems reasonable that pediatric anesthesiologists may have higher levels of P450 2E1 autoantibodies than general anesthesiologists as a consequence of direct exposure to inhaled anesthetics from mask inductions and uncuffed endotracheal tubes. To test this idea, and to determine whether anesthesiologists also have ERp58 autoantibodies in their sera, we collected sera from a large group of pediatric anesthesiologists at the 1998 Society of Pediatric Anesthesiology winter meeting and analyzed them for the presence of ERp58 and P450 2E1 autoantibodies. We found that pediatric anesthesiologists had higher levels of both ERp58 and P450 2E1 autoantibodies than did general anesthesiologists. However, because most anesthesiologists do not develop volatile anesthetic-induced liver injury, the results suggest that pathogenic ERp58 and P450 2E1 autoantibodies may not cause volatile anesthetic hepatitis.

Materials and Methods^
Chemicals and Reagents^
Affinity purified alkaline phosphatase (AP) conjugated goat antihuman immunoglobulin G was obtained from Life Technologies (Bethesda, MD), AP substrate reaction mixture was from BIO-RAD (Hercules, CA), and the alanine aminotransferase (ALT) test kit was from Sigma Diagnostics (St Louis, MO). Purified human ERp58 (12) and P450 2E1 (10) were prepared as previously described.

Human Sera^
The Joint Committee on Clinical Investigation IRB at the Johns Hopkins University School of Medicine approved the studies. After informed consent was obtained, sera from 105 pediatric anesthesiologists and 53 general anesthesiologists (including 3 nurse anesthetists) were collected and stored at -25°C until analyzed. A questionnaire was used to obtain demographic information from the subsets of individuals. Sera previously collected from patients with a clinical diagnosis of halothane hepatitis and control patients without exposure to volatile anesthetics were also analyzed (13). Because we were concerned that degradation of the proteins being tested may have occurred with some of the older sera from halothane hepatitis patients, as well as factitiously low autoantibodies in older control sera, all of the sera were reanalyzed, with the addition of new halothane hepatitis and control patients. We found that older sera results were analogous to new sera results in both groups.

Enzyme Linked Immunosorbent Assay of Human Sera^
All assays were performed in triplicate in the wells of Immulon® 4 microtiter plates (Dynex Incorporated, Chantilly, VA) in a total reaction volume of 100 µL by using a previously described method (14). Washings were performed with an Ultrawash Plus™ automated microplate washer (Dynatech Technologies). The test antigens were 0.5 µg of ERp58 or P450 2E1 in phosphate-buffered saline. AP product formation was determined at 405 nm after 60 and 90 min, respectively, using a SpectraMax 250 automatic plate reader (Molecular Devices, Sunnyvale, CA).

Demographics^
To determine whether differences in the autoantibody levels could be explained in terms of the baseline characteristics of each group, these characteristics were examined for both pediatric and general anesthesiologists. Autoantibody levels, age, work experience, and ALT levels were evaluated using Welch’s two-sided t-test for differences in means with unequal variances (15). Sex, history of autoimmune disease, liver disease, and infectious hepatitis were also examined in both groups using two-sided Fisher’s exact test for proportions (16).

ERp58 and P450 2E1 autoantibody levels for control and halothane hepatitis patients were compared with those levels for pediatric and general anesthesiologists without adjusting for other covariates. Box plots were constructed to compare the means and distributions of these levels among the four groups. The medians of the autoantibody levels of these groups were evaluated using pairwise Wilcoxon’s ranked sum tests (17).

Regression Models^
To determine whether differences in the autoantibody levels of pediatric and general anesthesiology groups were statistically different because of predictor variables, two separate multiple regression models were constructed (17). The two response variables, Y1 and Y2, were the ERp58 and P450 2E1 autoantibody levels, respectively. In both groups, autoantibody levels were skewed to the right. Therefore, these variables were transformed by natural logarithms in the regression models. Thus, log(Y1) and log(Y2) are linear functions of the predictor variables in such models. The candidate predictor variables included I1, type of anesthesiologist (0 for general, 1 for pediatric), I2, sex (0 for men, 1 for women), I3, history of autoimmune disease (0 for no, 1 for yes), I4, history of liver disease (0 for no, 1 for yes), and I5, history of infectious hepatitis (0 for no, 1 for yes). Other variables also included functions of age (yr), work experience (yr), and ALT levels (U/L), which were transformed by natural logarithms because they were skewed to the right. Additionally, because age and work experience were highly correlated, only one of these predictor variables was included in each model. Thus, the variables used in the model were X1 = log(age) (log-yr) or X2 = log(work experience) (log-yr) and X3 = log(ALT levels + 1) (log-IU/L).

Other characteristics obtained from the questionnaire included ethnicity, chronic medications, and alcohol consumption. None of these correlated with the autoantibody levels and were excluded from the final analysis. Additionally, the reported number of general anesthetics per day was also considered too unreliable to be included in the analysis. Interactions between anesthesiologist type (I1) and some of the other covariates were also considered. However, to avoid over-fitting the model, and because these covariates were balanced between both pediatric and general anesthesiologist groups, interactions between anesthesiologist type (I1) and the medical history indictor variables of autoimmune diseases (I3), liver disease (I4), and infectious hepatitis (I5) were excluded from the final models. One of the general anesthesiologists was deleted from the analysis because of missing questionnaire data.

All subset regression was used to select good candidate models. The criteria used to select the best model for each protein were Mallow’s Cp statistic, parsimony (i.e., models with the fewest number of variables), and interpretive value (18). The F test was used to determine whether the complete set of variables used in the model significantly explained the variability of the antibody levels. All variables included in the model were required to be significant at the 5% level. The statistical analyses were performed in the SAS 6.12 programming language (SAS Inc, Cary, NC).

Results^
Box-Plots Analyses^
Absorbance readings were linear with respect to antibody concentrations in the observed range of values. The pairwise Wilcoxon’s ranked sum test showed that the median of the ERp58 autoantibody levels for the control patients was significantly lower than the median levels for pediatric anesthesiologists and halothane hepatitis patients (Fig. 1). The median for the general anesthesiologist autoantibody levels was significantly lower than that for the pediatric anesthesiologists and halothane hepatitis patients, but it was not different from control patients. In contrast, the median autoantibody levels for the pediatric and halothane hepatitis patients could not be distinguished. Similar analysis showed that the median of the P450 2E1 autoantibody levels for control patients was significantly lower than the median levels for general anesthesiologists, pediatric anesthesiologists, and halothane hepatitis patients (Fig. 2). Furthermore, there was a highly significant increasing trend for the median levels of general anesthesiologists, pediatric anesthesiologists, and halothane hepatitis patients.


Figure 1. Box plots of the 58-kd endoplasmic reticulum protein (ERp58) autoantibody levels. The box plots were derived from the ELISA determinations of ERp58 serum autoantibody levels as expressed as absorbance at 405 nm on a logarithmic scale for controls (n = 20), general anesthesiologists (n = 52), pediatric anesthesiologists (n = 105), and halothane hepatitis patients (n = 20). The thick bar denotes the median, the lower and upper edges of the box denote the inter-quartile range (IQR), the whiskers extend to all observations within 1.5 times the IQR of the median, and the outliers are designated by asterisks. The shaded region denotes a 95% confidence interval for the median. The median of the ERp58 autoantibody levels for the control patients was significantly lower than the median level for pediatric anesthesiologists (P < 0.0001) and halothane hepatitis patients (P < 0.001). The median for the general anesthesiologists was significantly lower than the median for pediatric anesthesiologists and halothane hepatitis patients (P < 0.0001). In contrast, the median for the pediatric anesthesiologists and halothane hepatitis patients could not be distinguished (P = 0.19).

Figure 2. Box plots of cytochrome P450 2E1 (P450 2E1) autoantibody levels. The box plots were derived from the ELISA determinations of P450 2E1 serum autoantibody levels as expressed as absorbance at 405 nm on a logarithmic scale for controls (n = 20), general anesthesiologists (n = 52), pediatric anesthesiologists (n = 105), and halothane hepatitis patients (n = 20). The thick bar denotes the median, the lower and upper edges of the box denote the inter-quartile range (IQR), the whiskers extend to all observations within 1.5 times the IQR of the median, and outliers are designated by asterisks. The shaded region denotes a 95% confidence interval for the median. The median of the P450 2E1 autoantibody levels for control patients was significantly lower than the median level for general anesthesiologists (P = 0.0009), pediatric anesthesiologists, and halothane hepatitis patients (P < 0.0001). Furthermore, there was a highly significant increasing trend for the median levels of general anesthesiologists, pediatric anesthesiologists, and halothane hepatitis patients (general < pediatric < halothane hepatitis patients, P < 0.0001).
A scatter plot of the ERp58 autoantibody levels (y axis) plotted against the P450 2E1 autoantibody levels (x axis) also indicated that the pediatric and general anesthesiologists formed two distinct groups (Fig. 3). Although there was some overlap in the ERp58 levels between the two groups, all but a few of the pediatric anesthesiologists had higher P450 2E1 levels than the general anesthesiologists. This scatter plot emphasizes that pediatric anesthesiologists tend to have significantly higher levels of these autoantibodies than general anesthesiologists.


Figure 3. Plot of the 58-kd endoplasmic reticulum protein (ERp58) against cytochrome P450 2E1 (P450 2E1) autoantibody levels. The scatter plot was derived from the ELISA determination of ERp58 and P450 2E1 serum autoantibody levels as expressed as absorbance at 405 nm on a logarithmic scale. Solid symbols denote general anesthesiologists, open symbols denote pediatric anesthesiologists, squares denote men, and circles denote women. The dashed lines show the means of the autoantibody levels for the controls, whereas the dotted lines show the means ± 2 sd of the autoantibody levels for the controls.
Demographics of Pediatric and General Anesthesiologists^
A questionnaire was used to determine whether differences in ERp58 and P450 2E1 autoantibody levels between the two groups of anesthesiologists could be explained in terms of the baseline characteristics of each group. Three of the general anesthesiologists were certified registered nurses trained in anesthesiology, whereas the rest were physicians. For the purpose of simplicity, the nurse anesthetists were included in the general anesthesiology group because their practice was exclusive to adult patients. The two-sided Fisher’s exact test for proportions indicated that the number of women and history of autoimmune disease, liver disease, and infectious hepatitis were balanced between the two groups of anesthesiologists (Table 1). However, the Welch’s two-sided t-test for a difference in means, with unequal variances for the two groups of anesthesiologists, indicated that the means for age and work experience significantly differed (Table 1). Nevertheless, because there was a substantial overlap between these characteristics for the two groups, they were considered balanced for the purposes of this analysis. Furthermore, the autoantibody levels of pediatric anesthesiologists did not depend on age or work experience. The means for ALT levels did not differ between pediatric and general anesthesiologists and were within normal limits for persons without liver disease.


Table 1. Baseline Characteristics of Pediatric and General Anesthesiologists: Qualitative Comparisons of Pediatric and General AnesthesiologistsValues given are number (%) or mean ± sd.The Welch’s two-sided t-test for a difference in means with unequal variances indicated that the means for age and work experience differed between the two groups of anesthesiologists (P values = 0.002). The means for alanine aminotransferase (ALT) levels were not different between the two groups of anesthesiologists (P = 0.14). The two-sided Fisher’s exact test for proportions indicated that the number of women and history of autoimmune disease, liver disease, and infectious hepatitis were balanced between the two groups of anesthesiologists (P >0.2).
Regression Models^
Two separate multiple regression models were constructed for the ERp58 and P450 2E1 autoantibody levels on the basis of possible predictor variables possessed by the pediatric and general anesthesiologists. Three pediatric and one general anesthesiologist demonstrated very high ERp58 autoantibody levels. Although there were no distinguishing characteristics of this group, except that they were all male, linear regression analyses were performed both with and without these four values to determine their impact on the analyses. These autoantibody levels were judged to be accurate, but for the purposes of making inferences about most of these values, the outliers were deleted in the regression model for ERp58.

The best multiple regression model for ERp58 autoantibody levels with the four outliers removed (n = 153 observations) was MATH where expdenotes the natural exponential function, with an adjusted R2 score of 54.7%. The F test shows that the variables pediatric anesthesiologists and female sex explain why ERp58 autoantibody levels were higher among pediatric anesthesiologists than general anesthesiologists (Table 2). Specifically, pediatric anesthesiologists have higher ERp58 autoantibody levels than general anesthesiologists, and female anesthesiologists tend to have higher levels of autoantibodies to this protein than do male anesthesiologists with the same covariates. In the same way, the variable age of general anesthesiologists explains why some of their ERp58 autoantibody levels can be higher than those of other general anesthesiologists. Specifically, as age increases, general anesthesiologists tend to have higher levels of autoantibodies to ERp58. The other predictor variables did not contribute significantly to this model in the presence of the above covariates.


Equation U1

Table 2. Multiple Regression Model for ERp58 Autoantibody Levels.Y1 = exp {-2.96 + 1.76 I1 + 0.11 I2 + 0.35 (1 - I1)X1 + error}The F-test shows that the variables I1, pediatric anesthesiologists and I2, female sex explain why ERp58 autoantibody levels are higher among pediatric anesthesiologists than general anesthesiologists (P = 0.0001). Pediatric anesthesiologists have higher ERp58 autoantibody levels than general anesthesiologists (P = 0.001), and female anesthesiologists tend to have higher levels of autoantibodies to this protein than do male anesthesiologists with the same covariates (P = 0.004). As age increases, general anesthesiologists tend to have higher levels of autoantibodies to ERp58 (P = 0.02).
Similarly, the best multiple regression model for P450 2E1 autoantibody levels, based on n = 157 observations, was MATH with an adjusted R2 score of 70.0%. The F test shows that the variables pediatric anesthesiologists and female pediatric anesthesiologists explain why P450 2E1 autoantibody levels are higher in pediatric anesthesiologists than general anesthesiologists (Table 3). Specifically, pediatric anesthesiologists have higher levels of P450 2E1 autoantibodies than general anesthesiologists with the same covariates, and female pediatric anesthesiologists have higher levels of P450 2E1 autoantibodies than male pediatric anesthesiologists with the same covariates. Moreover, in an alternative model with simply sex (I2), rather than an interaction between sex and anesthesiologist type (I1I2), the coefficient for sex has a P value of 0.08. In the same way, ALT levels of general anesthesiologists explain why the P450 2E1 autoantibody levels of some general anesthesiologists can be higher than those of other general anesthesiologists. Specifically, those general anesthesiologists with higher ALT levels had higher levels of P450 2E1 autoantibodies. The other predictor variables did not contribute significantly to this model in the presence of the above covariates.


Equation U2

Table 3. Multiple Regression Model for P450 2E1 Antibody LevelsY2 = exp {-1.73 + 1.35 I1 + 0.12 I1I2 + 0.16 (1 - I1)X3 + error}.The F-test shows that the variables I1,I2, female pediatric anesthesiologists explain why cytochrome P450 2E1 autoantibody levels are higher in pediatric anesthesiologists than general anesthesiologists (P = 0.0001). Pediatric anesthesiologists have higher levels of P450 2E1 autoantibodies than general anesthesiologists with the same covariates (P = 0.0001), and female pediatric anesthesiologists have higher levels of P450 2E1 autoantibodies than male pediatric anesthesiologists with the same covariates (P = 0.04). General anesthesiologists with higher alanine aminotransferase (ALT) levels had higher levels of P450 2E1 autoantibodies (P = 0.05).
Of all of the anesthesiologists studied, only one female pediatric anesthesiologist had symptoms of recurrent hepatitis after exposure to volatile anesthetics. These symptoms resolved after the she was removed from the operating room environment. ERp58 and P450 2E1 autoantibody levels were 0.382 and 1.253, respectively, both of which were well within the increased values seen for pediatric anesthesiologists without active liver disease.

Discussion^
The present study has shown that pediatric anesthesiologists have significantly higher serum levels of anesthetic hepatitis-associated ERp58 and P450 2E1 autoantibodies than general anesthesiologists and control patients, whereas general anesthesiologists only have increased serum levels of P450 2E1 autoantibodies when compared with the control patients. These results suggest that chronic occupational exposure to volatile anesthetics can lead to continuous hepatic metabolism of volatile anesthetics and formation of immunogenic TFA protein adducts, including TFA-ERp58 and TFA-P450 2E1, which can induce formation of the ERp58 and P450 2E1 autoantibodies associated with volatile anesthetic hepatitis (7). The results have also shown that female anesthesiologists had higher levels of ERp58 autoantibodies than male anesthesiologists, whereas female pediatric anesthesiologists had higher levels of P450 2E1 autoantibodies than all other anesthesiologists. Antibody responses after immunization are often greater in women than in men and are believed to be caused either by estrogens or pituitary hormones (19). Whether these hormones have a similar effect in the present study is not known.

Only one female pediatric anesthesiologist developed liver injury, even when all of the pediatric anesthesiologists as a group had increased serum levels of anesthetic hepatitis-associated ERp58 and P450 2E1 autoantibodies that were not significantly different from those of halothane hepatitis patients. Thus, our findings suggest that ERp58 and P450 2E1 autoantibodies may not have a role in the development of inhaled anesthetic hepatitis. Alternatively, it is possible that antigen-specific cytotoxic T cells, instead of autoantibodies, which are directed against peptides derived from ERp58 and P450 2E1, may cause volatile anesthetic hepatitis. In this regard, antigen-specific cytotoxic T cells, but not autoantibodies, appear to have a role in the pathogenesis of liver injury caused by hepatitis B infection (20).

In contrast, humoral reactions, cellular immune reactions, or both against native ERp58 and P450 2E1 may not have a role in volatile anesthetic-induced hepatitis. Perhaps only TFA-altered forms of these autoantigens can be targets of pathogenic antibodies or cytotoxic T cells. In this regard, previous studies have demonstrated that only halothane hepatitis patients, but not patients exposed to halothane who did not develop hepatitis, or patients with other forms of liver disease, have serum antibodies that react with TFA liver microsomal antigens (7). Furthermore, other cellular targets of the reactive acyl halide metabolites of volatile anesthetics that also become TFA modified, such as a carboxylesterase, protein disulfide isomerase, ERp72, and glucose-related proteins 78 and 94, could potentially become the immunogens that lead to volatile anesthetic-induced hepatitis (7).

In conclusion, we found significantly higher levels of ERp58 and P450 2E1 serum autoantibodies in pediatric anesthesiologists when compared with general anesthesiologists. The mathematical regression models verify that female anesthesiologists, both pediatric and general, have higher levels of autoantibodies to ERp58 than male anesthesiologists, and female pediatric anesthesiologists tend to have higher levels of P450 2E1 autoantibodies than male pediatric anesthesiologists. Still, only one of the female pediatric anesthesiologists developed symptoms of anesthetic hepatitis, even though these autoantibodies have been associated with volatile anesthetic-induced hepatitis. These findings suggest that ERp58 and P450 2E1 serum autoantibodies may not have a role in volatile anesthetic-induced hepatitis and that other immune mechanisms may determine whether halogenated volatile anesthetic-induced hepatitis occurs.

References^
1. Klatskin G, Kimberg DV. Recurrent hepatitis attributable to halothane sensitization in an anesthetist. N Engl J Med 1969; 280: 515–22. [Medline Link] [Context Link]

2. Neuberger J, Vergani D, Mieli-Vergani G, et al. Hepatic damage after exposure to halothane in medical personnel. Br J Anaesth 1981; 53: 1173–7. [Medline Link] [Context Link]

3. Keiding S, Dossing M, Hardt F. A nurse with liver injury associated with occupational exposure to halothane in a recovery unit. Dan Med Bull 1984; 31: 255–6. [Medline Link] [Context Link]

4. Sutherland DE, Smith WA. Chemical hepatitis associated with occupational exposure to halothane in a research laboratory. Vet Hum Toxicol 1992; 34: 423–4. [Medline Link] [Context Link]

5. Whitcher CE, Cohen EN, Trudell JR. Chronic exposure to anesthetic gases in the operating room. Anesthesiology 1971; 35: 348–53. [Medline Link] [Context Link]

6. Njoku D, Laster MJ, Gong DH, et al. Biotransformation of halothane, enflurane, isoflurane, and desflurane to trifluoroacetylated liver proteins: association between protein acylation and hepatic injury. Anesth Analg 1997; 84: 173–8. [Fulltext Link] [Medline Link] [BIOSIS Previews Link] [Context Link]

7. Pohl LR, Pumford NR, Martin JL. Mechanisms, chemical structures and drug metabolism. Eur J Haematol Suppl 1996; 60: 98–104. [Medline Link] [Context Link]

8. Martin JL, Pumford NR, LaRosa AC, et al. A metabolite of halothane covalently binds to an endoplasmic reticulum protein that is highly homologous to phosphatidylinositol-specific phospholipase C-alpha but has no activity. Biochem Biophys Res Commun 1991; 178: 679–85. [Medline Link] [Context Link]

9. Martin JL, Reed GF, Pohl LR. Association of anti-58 kDa endoplasmic reticulum antibodies with halothane hepatitis. Biochem Pharmacol 1993; 46: 1247–50. [Medline Link] [BIOSIS Previews Link] [Context Link]

10. Bourdi M, Chen W, Peter RM, et al. Human cytochrome P450 2E1 is a major autoantigen associated with halothane hepatitis. Chem Res Toxicol 1996; 9: 1159–66. [Medline Link] [BIOSIS Previews Link] [Context Link]

11. Eliasson E, Kenna JG. Cytochrome P450 2E1 is a cell surface autoantigen in halothane hepatitis. Mol Pharmacol 1996; 50: 573–82. [Medline Link] [BIOSIS Previews Link] [Context Link]

12. Bourdi M, Demady D, Martin JL, et al. cDNA cloning and baculovirus expression of the human liver endoplasmic reticulum P58: characterization as a protein disulfide isomerase isoform, but not as a protease or a carnitine acyltransferase. Arch Biochem Biophys 1995; 323: 397–403. [Medline Link] [BIOSIS Previews Link] [Context Link]

13. Martin JL, Kenna JG, Pohl LR. Antibody assays for the detection of patients sensitized to halothane. Anesth Analg 1990; 70: 154–9. [Medline Link] [Context Link]

14. Njoku DB, Pohl LR, Sokoloski EA, et al. Immunochemical evidence against the involvement of cysteine conjugate beta-lyase in compound A nephrotoxicity in rats. Anesthesiology 1999; 90: 458–69. [Fulltext Link] [Medline Link] [BIOSIS Previews Link] [Context Link]

15. Gaylor DW. Satterthwaite’s formula. In: Kotz S, Johnson NL. Encyclopedia of Statistical Sciences, Vol 8. New York: John Wiley and Sons, Inc, 1988: 216–62. [Context Link]

16. Agresti A. Categorical data analysis. New York: John Wiley & Sons, Inc, 1990: 59–67. [Context Link]

17. Altman DG. Practical statistics for medical research. New York: Chapman and Hall, 1991: 336–9. [Context Link]

18. Draper NR, Smith H. Applied regression analysis. 2nd ed. New York: John Wiley & Sons, Inc, 1981:296–305, 93–4. [Context Link]

19. Whitacre CC, Reingold SC, O’Looney PA. A gender gap in autoimmunity. Science 1999; 283: 1277–8. [Medline Link] [BIOSIS Previews Link] [Context Link]

20. Koziel MJ. The immunopathogenesis of HBV infection. Antivir Ther 1998; 3S: 13–24. [Medline Link] [Context Link]


Accession Number: 00000539-200202000-00003
Copyright (c) 2000-2002 Ovid Technologies, Inc.
Version: rel4.6.0, SourceID 1.5852.1.103