Nonalcoholic fatty liver disease (NAFLD) is the most common cause of liver disease in children. NAFLD is a clinicopathological diagnosis characterized by the accumulation of macrovesicular fat in hepatocytes in the absence of alcohol consumption. NAFLD encompasses a spectrum of histopathological features ranging from simple steatosis to steatosis with inflammation, ballooning degeneration, and pericellular fibrosis (nonalcoholic steatohepatitis, NASH) to cirrhosis. The existing population-based prevalence studies suggest that NAFLD is a global problem with reports published in North and South America, Europe, Australia, and Asia (1). Because NAFLD is diagnosed by liver biopsy, it is hard to estimate the prevalence in children in a population-based study. In a community representative autopsy study (2), which was based on liver histology conducted from 1993 to 2003, the standardized prevalence of fatty liver disease in children ages 2 to 19 years was estimated at 9.6% (2).
The detailed effects of intrahepatocellular lipid accumulation on hepatic function are poorly understood. In particular, knowledge of the potential consequences of steatosis on the drug-metabolizing capability of the human liver is limited. Several studies have reported on cytochrome P450 (CYP)–mediated drug metabolism in animal models of fatty liver disease; however, there are few studies evaluating the activity of UDP-glucuronyltransferases (UGTs) and other phase II enzymes in NAFLD. Microsomal CYP dysregulation has been well documented in animal models of steatosis, obesity, and steatohepatitis (3,4). Drug clearance may also be impaired in NAFLD. In obesity, which is highly associated with NAFLD, changes in pharmacokinetics are frequently observed as a result of altered drug distribution and biotransformation (5). Lipid accumulation of drugs in peripheral adipose tissue is known to increase the volume of drug distribution and decrease drug clearance. It is conceivable that drugs may also accumulate in the lipid-filled subcellular compartments of the liver in NAFLD and similarly limit the rates of drug elimination. With diminished ability to metabolize and eliminate commonly used medications and dietary supplements, children with fatty liver disease may be at increased risk for drug-induced hepatotoxicity. In children with diagnosed NAFLD, lower metabolic activity could represent a hazard because drugs given in conventional doses may accumulate and produce toxicity. In subjects with obesity with undiagnosed NAFLD, treatment with potentially hepatotoxic medications metabolized through these pathways may cause idiopathic drug-induced liver injury.
Acetaminophen (APAP) is the most common cause of acute liver failure in the pediatric population (6). Although safe at therapeutic doses, overdoses can lead to severe hepatotoxicity, especially in individuals with preexisting liver disease (7). In APAP overdose, a highly reactive metabolite is formed and covalently binds to macromolecules to cause cellular damage (6). Recent studies (8) suggest that steatohepatitis sensitizes the liver to APAP toxicity in mice. The purpose of the present study was to evaluate UGT activity and APAP pharmacokinetics in children with fatty liver disease, with the hypothesis that potentially disordered hepatic metabolism may alter drug disposition and predispose to idiosyncratic hepatic toxicity.
PATIENTS AND METHODS
Participants
Inclusion criteria for cases were boys ages 10 to 17 years with histological evidence of NAFLD based on liver biopsy obtained between January 2008 and November 2009 at the University of California, San Diego (UCSD). Given that the present study is a pilot, only male subjects were chosen to minimize other potentially confounding variables associated with sex. NAFLD is more common in boys, even when matched for other variables. At UCSD 80% of children with NAFLD are boys. Table 1 summarizes the basic characteristics of the study subjects. Inclusion criteria for controls were serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) <30 U/L and magnetic resonance (MR) spectroscopy with <5% calculated liver fat, who were otherwise matched to the children with NAFLD for age and sex. MR spectroscopy was performed on a 3-T MR scanner at the Department of Radiology at UCSD by an MR physicist with 10 years' experience in liver spectroscopy.
;)
TABLE 1: Demographic characteristics of study groups
Participants were excluded from study consideration if any of the following criteria were present: other causes of chronic hepatitis including hepatitis B, hepatitis C, α1-antitrypsin deficiency, autoimmune hepatitis, Wilson disease, drug toxicity, total parenteral nutrition, and chronic alcohol intake; history of significant renal, gastrointestinal (other than NAFLD), or cardiac disease; taking medications that induce or inhibit drug-metabolizing enzyme activity, including smoking tobacco products; inability to swallow medication; allergy or intolerance to APAP.
Study Protocol
Following an overnight fast and a 3-day drug-free baseline period, each subject received a single oral dose of APAP 5 mg/kg (up to 325 mg). In addition, subjects received simultaneous administration of a single oral dose of dextromethorphan 0.3 mg/kg (up to 15 mg) and 8 oz Diet Coke (Coca-Cola Company, Atlanta, GA) as in vivo probes to study the enzymes CYP2D6, CYP1A2, and CYP3A (reported elsewhere). Food was withheld and subjects remained in the upright position for 2 hours following drug administration. Baseline urine and saliva samples were obtained before drug administration. Total voided urine was collected 0 to 4 and 4 to 24 hours following drug administration, and nonstimulated saliva samples were collected at intervals for up to 4 hours (15, 30, 45, 60, 120, and 240 minutes). A single blood sample was collected 4 hours after drug administration. All of the samples (except the 4- to 24-hour urine samples that were collected at home) were collected in the UCSD General Clinical Research Center. The harvested plasma, urine, and saliva aliquots were stored frozen at −70°C until analyzed. The study was approved by the UCSD institutional review board (#080430). Written informed consent and assent were obtained from subjects' parents and subjects, respectively.
Analysis of APAP and Metabolites
APAP and APAP-glucuronide (APAP-G) metabolites were measured by high-performance liquid chromatography (HPLC). Orthophosphoric acid 85%, perchloric acid 70%, acetonitrile HPLC grade, and HPLC-grade water were purchased from Fisher Scientific (Fair Lawn, NJ). Potassium hydroxide 50% was purchased from RICCA Chemical Co (Arlington, TX). Formic acid 88%, APAP, and APAP-G were purchased from Sigma Aldrich (St Louis, MO). Normal human ethylenediaminetetraacetic acid plasma and normal human saliva were purchased from Biochemed (Winchester, VA). Normal human urine was obtained from a normal donor. The HPLC system consisted of an autosampler (SpectraSYSTEM AS3000), HPLC pump (SpectraSYSTEM P4000), a detector (Spectra Focus Forward Optical Scanner), a data integrator (ChromQuest version 4.0, Thermo Electron, San Jose, CA), and a C18 reversed-phase HPLC column (MAC-MOD ACE 5, 4.6 × 150 mm). All of the assays were conducted with the analytical column at ambient temperature.
Standard curves were prepared using standard solutions of APAP and APAP-G with normal human urine, plasma, or saliva. The stock solution was serially diluted with normal human urine, plasma, or saliva to produce the required range of concentrations. Calibration standards were evaluated using a least-squares linear regression algorithm to plot the peak height versus concentration with 1/response weighting. Standard curves were automatically generated by ChromQuest software (San Jose, CA).
Data Analysis
The activity of UGT was estimated by the plasma ratio of APAP-G to APAP at 4 hours. APAP elimination rate constant (kel/h) over the linear range was determined by linear regression of the logarithm of the concentration against time. It involved at least 5 sampling points. The half-life was obtained from the elimination rate constant (t1/2 = ln 2/kel). A 1-compartment model was used to describe APAP elimination (9). The area under the concentration/time curves from 0 to 4 hours was calculated using PRISM 5.0 software (GraphPad Software Inc, La Jolla, CA). The clearance of APAP was calculated by dividing the initial APAP dose by the area under the concentration/time curve for each patient and expressed per kilogram of body weight.
Statistical Analysis
Data are expressed as mean ± SD. Comparison between NAFLD and control at a single time point was made using the Wilcoxon rank sum test. The comparison between different time points for each group was made by 2-way analysis of variance followed by the Bonferroni post hoc test. Statistical analyses were performed using PRISM 5.0 (GraphPad Software Inc). Two-sided P values <0.05 were considered statistically significant.
RESULTS
Subject Characteristics
Twenty-four children were included in the present study, 12 children with NAFLD and 12 children without NAFLD. There was no significant difference between the groups with respect to age. BMI was significantly higher in subjects with NAFLD (34.0 ± 6.1) as compared with controls (26.2 ± 10.9, P = 0.043). As expected, plasma aminotransferases were substantially and significantly elevated in subjects with NAFLD as compared with controls (P < 0.0018). Serum ALT was most dramatically elevated in subjects with NAFLD (100 ± 73) as compared with controls (20 ± 8, P = 0.001). Of the patients with NAFLD, 42% (n = 5) had a fibrosis stage of 2 or greater on liver biopsy.
APAP Pharmacokinetics
The mean dose of APAP administered to children with NAFLD was 3.6 mg/kg (STD 0.8), and the mean dose administered to children in the control group was 3.8 mg/kg (STD 0.9). Salivary concentrations were used to define the pharmacokinetic profile of APAP in children with and without fatty liver disease (10). Figure 1 shows the concentration of APAP in saliva for 4 hours in boys with NAFLD compared with controls. No APAP metabolites were detectable in the saliva of either group. Figure 2 shows the correlation between plasma and salivary APAP concentrations in all of the subjects at 4 hours. A regression of the concentration of APAP in saliva (y) versus the plasma concentration (x) showed a significant linear relation (r2 = 0.595, P < 0.0001), which could be described by the equation y = 1.07x + 0.413. Table 2 summarizes selected APAP pharmacokinetic parameters. Nonparametric analyses failed to show a significant difference in any of the pharmacokinetic parameters (clearance, half-life, area under the curve, and peak APAP concentration) in children with NAFLD compared with controls.
FIGURE 1: APAP concentrations in saliva at various time points after administration of a single-dose APAP in children with NAFLD compared with controls. No APAP metabolites were measured in saliva. Values are the mean ± SD. APAP = acetaminophen; NAFLD = nonalcoholic fatty liver disease.
FIGURE 2: Scatterplot depicting correlation between saliva and plasma concentrations of acetaminophen in all 24 children at 4 hours. Linear regression showed a significant linear relation (r 2 = 0.595, P < 0.0001). Each point on the graph represents 1 study participant. APAP, see Figure 1 caption.
TABLE 2: Acetaminophen pharmacokinetic data calculated from saliva samples (reported as median, range values)
Plasma Concentration APAP and APAP-G
Figure 3 shows the concentration of APAP and APAP-G in plasma at 4 hours. The concentration of APAP-G was 35% higher in children with NAFLD compared with controls (P = 0.0071). There was a statistically significant increase in the ratio of APAP-G to APAP (P = 0.0277) in children with NAFLD.
FIGURE 3: APAP and APAP-G concentrations in serum 4 hours after a single dose of APAP in children with NAFLD compared with control. Values are the mean ± SD. *, Significant difference from the control group (P < 0.05). APAP and NAFLD, see Figure 1 caption; APAP-G = acetaminophen-glucuronide.
Excretion APAP and APAP-G in Urine
Figure 4 shows the cumulative concentration of APAP-G excreted in urine samples collected during 24 hours. The pooled concentration of APAP-G in urine collected 4 hours after APAP administration was 2.2-fold higher in children with NAFLD compared with children without liver disease (P = 0.0151). The pooled concentration of APAP-G in urine collected between 4 and 24 hours after APAP administration was 2.0-fold higher in children with NAFLD (P = 0.0210).
FIGURE 4: Cumulative concentrations of APAP-G in the urine at 0 to 4 hours and 4 to 24 hours after APAP administration. Values are the mean ± SD. *, Significant difference from the control group (P < 0.05). APAP-G and NAFLD, see Figure 1 caption.
DISCUSSION
This is the first study to our knowledge to assess drug metabolism in children with NAFLD. We show elevated serum and urine concentrations of APAP-G in children with NAFLD. However, salivary time-concentration profiles after a single dose of APAP in children with NAFLD appear to be the same in children with fatty liver disease as in children without liver disease. The peak serum APAP concentrations, APAP half-life, and APAP clearance were nearly identical in these groups. The altered metabolism of APAP in children with NAFLD does not appear to affect its rate of elimination. These data suggest that a similar dosage schedule should apply for children with NAFLD as for normal children.
The safety of APAP in children with NAFLD is more difficult to assess. Hepatotoxicity is dependent on the balance between the rate of the toxic metabolite (N-acetyl-p-benzoquinone imine, NAPQI) formation through oxidative pathways, the capacity of the safe elimination pathways of sulfate and glucuronide conjugation, and the rate of hepatic glutathione synthesis. These processes are largely undefined in children with NAFLD.
In normal subjects, 90% of APAP is metabolized to APAP-G and APAP-sulfate (APAP-S) in the liver by UGT and sulfotransferases and subsequently excreted in urine (6). Sulfate conjugation is the dominant metabolic pathway in neonates and young children, but glucuronide metabolism increases with age (10–12). Less than 10% of APAP is oxidized by the CYP pathway (particularly CYP2E1) to NAPQI, which is immediately neutralized by conjugation with glutathione and excreted in the urine as cysteine and mercapturic acid conjugates (APAP-C and APAP-M) (6). However, this pathway assumes greater significance in conditions in which glucuronidation is reduced or CYP2E1 activity is increased (6).
CYP2E1 activity is known to be upregulated in clinical settings associated with NASH, particularly diabetes and obesity (13–15), and has been shown in adults with NASH (15–17) and animal models of steatohepatitis (18). In fact, enhanced CYP2E1 activity is associated with hepatic microsomal lipid peroxidation and oxidative stress in NASH and is believed to play a role in the progression from simple steatosis to NASH (4,19). Enhanced CYP2E1 activity in NASH could predispose the liver to APAP-induced hepatotoxicity. Recently, it was shown that steatohepatitis sensitizes the liver to APAP-induced hepatotoxicity and hepatic failure in mice (20). In the present study, the contribution of the CYP2E1 oxidative pathway to APAP metabolism in children with NAFLD was not addressed. Urine APAP-C and APAP-M concentrations reflect the amount of APAP converted to the reactive intermediate NAPQI by CYP oxidation (6). A significant elevation in these metabolites may suggest that children with NAFLD are at increased risk for APAP-induced hepatotoxicity. In this study, however, measurement of these toxic adducts was not feasible in light of a single subtherapeutic dose of APAP.
Here we show a significant increase in the formation of APAP-G in children with fatty liver disease, which is likely due to UGT upregulation. Although this may represent an overall increase in APAP metabolism, it is far more likely a reflection of decreased activity in other metabolic pathways not measured in the present study. This hypothesis is supported by the normal pharmacokinetic parameters, particularly area under the curve and peak concentration, in the face of elevated APAP-G formation. We suspect that the higher rate of APAP-G formation is compensating for a deficiency in APAP-S formation in children with NAFLD. Impaired sulfate formation in NAFLD could present a significant problem for APAP metabolism in young children in whom conjugation to sulfate is the predominant elimination pathway. Because sulfate conjugation is the dominant metabolic pathway in normal children and potentially impaired in children with NAFLD, a limitation of the present study was the inability of our assays to quantify APAP-S in urine. In a study of APAP metabolism in children ages 6 months to 7 years, chronic liver disease of varying etiologies (biliary atresia, Alagille syndrome, total parenteral nutrition cholestasis, α1-antitrypsin deficiency, chronic active hepatitis, and congenital hepatic fibrosis) was associated with an increased urinary APAP-G to APAP-S ratio (9,21). Because the present study did not include a group of children with chronic liver diseases other than NAFLD, it is impossible to determine whether the increased formation of APAP-G we observed was specific to NAFLD as opposed to a more general effect of liver disease in children. We also cannot exclude the possibility of a reverse association between UGT activity and the development of NAFLD—it is conceivable that children who metabolize APAP toward APAP-G are more prone to develop NAFLD. Regardless of the nature of the metabolites formed, the present study shows that the pharmacokinetics of APAP in children with NAFLD are the same as in children without liver disease.
Additionally, we demonstrate salivary sampling as a viable alternative to plasma sampling for defining APAP pharmacokinetic parameters in children. No measurable levels of APAP conjugates were detectable in the saliva of our patients. However, a highly significant relation between plasma and salivary concentrations of APAP was observed, and the estimated pharmacokinetic parameters were consistent with published values (9,21). The use of saliva for drug monitoring and future pharmacokinetic studies in children would offer a significant advantage because repetitive venipuncture can be associated with anxiety and trauma in children.
Of note, the children in the present study were given a cocktail of APAP, dextromethorphan, and caffeine for the simultaneous measurement of the activity of multiple drug-metabolizing enzymes (CYP2D6, CYP1A2, CYP3A4, and UGT). The use of such cocktails has been validated in several studies (22,23), which indicate no pharmacokinetic or pharmacodynamic interactions between the probe drugs.
In conclusion, our data suggest that APAP metabolism is altered in children with fatty liver disease. These findings support the hypothesis that NAFLD has a significant effect on drug metabolism and underscore the importance of evaluating in detail the potential influence of hepatic steatosis on the metabolism, safety, and efficacy of potentially hepatotoxic drugs. Here we show that regardless of the metabolites formed, the pharmacokinetics of a single 5 mg/kg dose of APAP is the same in children with NAFLD as in children with normal liver function. These findings suggest that the pharmacotherapeutic dosing regimen for APAP in children with NAFLD should be the same as for children without liver disease. Future studies measuring APAP-C and APAP-M formation following exposure to the recommended doses of APAP are necessary to determine whether children with NAFLD are at an increased risk for APAP-induced hepatic injury.
Acknowledgments
The authors wish to thank Steven Rossi and Rowena Espina for performing drug assays.
REFERENCES
1. Barshop NJ, Sirlin CB, Schwimmer JB,
et al. Review article: epidemiology, pathogenesis and potential treatments of paediatric non-alcoholic fatty liver disease. Aliment Pharmacol Ther 2008; 28:13–24.
2. Schwimmer JB, Deutsch R, Kahen T,
et al. Prevalence of fatty liver in children and adolescents. Pediatrics 2006; 118:1388–1393.
3. Leclercq I, Horsmans Y, Desager JP,
et al. Reduction in hepatic cytochrome P-450 is correlated to the degree of liver fat content in animal models of steatosis in the absence of inflammation. J Hepatol 1998; 28:410–416.
4. Gomez-Lechon MJ, Jover R, Donato MT. Cytochrome P450 and steatosis.
Curr Drug Metab 2009;10:692–9.
5. Cheymol G. Effects of obesity on pharmacokinetics implications for drug therapy. Clin Pharmacokinet 2000; 39:215–231.
6. Chun LJ, Tong MJ, Busuttil RW,
et al. Acetaminophen hepatotoxicity and acute liver failure. J Clin Gastroenterol 2009; 43:342–349.
7. Benson GD, Koff RS, Tolman KG. The therapeutic use of acetaminophen in patients with liver disease. Am J Ther 2005; 12:133–141.
8. Osabe M, Sugatani J, Fukuyama T,
et al. Expression of hepatic UDP-glucuronosyltransferase 1A1 and 1A6 correlated with increased expression of the nuclear constitutive androstane receptor and peroxisome proliferator-activated receptor alpha in male rats fed a high-fat and high-sucrose diet. Drug Metab Dispos 2008; 36:294–302.
9. Alam SN, Roberts RJ, Fischer LJ. Age-related differences in salicylamide and acetaminophen conjugation in man. J Pediatr 1977; 90:130–135.
10. Al-Obaidy SS, Li Wan Po A, McKiernan PJ,
et al. Assay of paracetamol and its metabolites in urine, plasma and saliva of children with chronic liver disease. J Pharm Biomed Anal 1995; 13:1033–1039.
11. Cummings AJ, King ML, Martin BK. A kinetic study of drug elimination: the excretion of paracetamol and its metabolites in man. Br J Pharmacol Chemother 1967; 29:150–157.
12. van der Marel CD, Anderson BJ, van Lingen RA,
et al. Paracetamol and metabolite pharmacokinetics in infants. Eur J Clin Pharmacol 2003; 59:243–251.
13. Raucy JL, Lasker JM, Kraner JC,
et al. Induction of cytochrome P450IIE1 in the obese overfed rat. Mol Pharmacol 1991; 39:275–280.
14. Dong ZG, Hong JY, Ma QA,
et al. Mechanism of induction of cytochrome P-450ac (P-450j) in chemically induced and spontaneously diabetic rats. Arch Biochem Biophys 1988; 263:29–35.
15. Emery MG, Fisher JM, Chien JY,
et al. CYP2E1 activity before and after weight loss in morbidly obese subjects with nonalcoholic fatty liver disease. Hepatology (Baltimore) 2003; 38:428–435.
16. Chalasani N, Gorski JC, Asghar MS,
et al. Hepatic cytochrome P450 2E1 activity in nondiabetic patients with nonalcoholic steatohepatitis. Hepatology (Baltimore) 2003; 37:544–550.
17. Weltman MD, Farrell GC, Hall P,
et al. Hepatic cytochrome P450 2E1 is increased in patients with nonalcoholic steatohepatitis. Hepatology (Baltimore) 1998; 27:128–133.
18. Weltman MD, Farrell GC, Liddle C. Increased hepatocyte CYP2E1 expression in a rat nutritional model of hepatic steatosis with inflammation. Gastroenterology 1996; 111:1645–1653.
19. Villanova N, Moscatiello S, Ramilli S,
et al. Endothelial dysfunction and cardiovascular risk profile in nonalcoholic fatty liver disease. Hepatology (Baltimore) 2005; 42:473–480.
20. Donthamsetty S, Bhave VS, Mitra MS,
et al. Nonalcoholic steatohepatitic (NASH) mice are protected from higher hepatotoxicity of acetaminophen upon induction of PPARalpha with clofibrate. Toxicol Appl Pharmacol 2008; 230:327–337.
21. Al-Obaidy SS, McKiernan PJ, Li Wan Po A,
et al. Metabolism of paracetamol in children with chronic liver disease. Eur J Clin Pharmacol 1996; 50:69–76.
22. Evans WE, Relling MV, Petros WP,
et al. Dextromethorphan and caffeine as probes for simultaneous determination of debrisoquin-oxidation and N-acetylation phenotypes in children. Clin Pharmacol Ther 1989; 45:568–573.
23. Streetman DS, Bleakley JF, Kim JS,
et al. Combined phenotypic assessment of CYP1A2, CYP2C19, CYP2D6, CYP3A, N-acetyltransferase-2, and xanthine oxidase with the “Cooperstown cocktail”. Clin Pharmacol Ther 2000; 68:375–383.
