|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 26, No 12 (April 20), 2008: pp. 1932-1939 © 2008 American Society of Clinical Oncology. DOI: 10.1200/JCO.2007.13.8404 Asparaginase May Influence Dexamethasone Pharmacokinetics in Acute Lymphoblastic Leukemia
From the Departments of Pharmaceutical Sciences, Oncology, and Biostatistics, St Jude Children's Research Hospital; and the Colleges of Medicine and Pharmacy, University of Tennessee, Memphis, TN Corresponding author: Mary V. Relling, PharmD, Department of Pharmaceutical Sciences, St Jude Children's Research Hospital, 332 N Lauderdale, Memphis, TN 38105-2794; e-mail: mary.relling{at}stjude.org
Purpose Dexamethasone is used widely in oncology, but pharmacokinetic studies are lacking. We evaluated dexamethasone pharmacokinetics in children with acute lymphoblastic leukemia. Patients and Methods We assessed 214 children with acute lymphoblastic leukemia who received 418 courses of oral dexamethasone (8 mg/m2/d) on days 1 and 8 of reinduction. Extensive asparaginase use preceded reinduction in the 101 children in the standard/high-risk treatment arm but not in the 113 children in the low-risk treatment arm. A one-compartment model with first-order absorption and disposition was fit to dexamethasone plasma concentrations by using maximum a posteriori probability estimation; we evaluated covariates by using linear mixed models. Results Interpatient and intrapatient variabilities in apparent clearance were substantial; they were 46% and 53%, respectively. Variability was explained by the serum albumin concentration (P < .0001), concomitant use of fentanyl (P = .008) and ketoconazole (P = .03), and age (P = .006). Apparent clearance was higher in the low-risk arm (P < .001) and was related to a greater serum albumin concentration (P < .001) and to a lower exposure to asparaginase than in the standard/high-risk arm. Hypoalbuminemia, a biomarker of asparaginase activity, was associated with a lower dexamethasone apparent clearance (P = .04) in patients in the standard/high-risk arm that was more pronounced in those not allergic to asparaginase. Ethnicity or gender did not explain apparent clearance variability. Conclusion Dexamethasone pharmacokinetics are highly variable and are related to the concurrent use of particular drugs, age, and treatment intensity. Patients allergic to asparaginase may be doubly disadvantaged: they not only suffer from diminished exposure to asparaginase but also, by maintaining high clearance of dexamethasone, may experience fewer antileukemic effects of dexamethasone.
Glucocorticoids are used extensively in adult and pediatric oncology as antileukemic,1,2 anti-inflammatory,3 and antiemetic agents.4-6 Although dexamethasone is useful in many settings, including in the treatment of childhood acute lymphoblastic leukemia (ALL), adverse events may be dose-limiting.7 In other patient populations (eg, organ transplant recipients), both desired and adverse effects are dose-related and may be related to pharmacokinetic variability, but there are no prior pharmacokinetic studies of dexamethasone in patients with ALL, and there are few studies in adult patients with cancer.8,9 To our knowledge, only a single pharmacokinetic study of dexamethasone in children (without ALL) has been reported,10 which showed a high variation in pharmacokinetics. Because of the limited nature of prior studies, the extent of inter- and intrapatient variability in dexamethasone pharmacokinetics among cancer patients, as well as the covariates for such variability, remains unclear. Our objectives were to characterize the pharmacokinetic parameters of dexamethasone among children with ALL in a controlled trial, to estimate intrapatient and interpatient variability in the systemic exposure to the drug, and to explore the contribution of covariates to variability in dexamethasone pharmacokinetics.
Patients We studied 214 children with ALL who were treated on protocol Total XV at St Jude Children's Research Hospital from July 2000 until December 2005.11 The institutional review board approved the study, and informed consent was obtained from parents/guardians or from patients. Patients were assigned to one of three risk categories: low, standard, or high risk, as described previously.11 Younger children were more likely to be assigned to the low-risk group, whereas older children were more likely to be assigned to the standard- or high-risk groups. Asparaginase allergy was graded by using the National Cancer Institute Common Toxicity Criteria version 2.0, and patients were categorized as those with no toxicity (grade 0) and those with grade 1 to 4 allergy before week 7 of continuation therapy (start of reinduction I).
Treatment Regimen and Sample Collection
Dexamethasone 8 mg/m2/d was given orally as tablets and was divided equally into three doses per day during days 1 through 8 of reinduction. The concomitant use of other drugs in the 72 hours before and 24 hours after dexamethasone was recorded by a pharmacokinetics research nurse (Appendix Table A1, online only). Blood (3 mL) was drawn into heparin-containing tubes before and 1, 2, 4, and 8 hours after the morning dexamethasone dose on days 1 and 8 of reinduction I, which corresponded to weeks 7 and 8 of continuation treatment, respectively. Plasma was stored at –80°C.
Dexamethasone Concentrations
Pharmacokinetic Model
The model-fitted curve for each patient was used to estimate the area under the concentration-time curve (AUC) from time 0 to 8 hours (AUC0
The population pharmacokinetics were determined using a two-stage approach.13 In the first stage, the pharmacokinetic parameters for each individual course were estimated as above. The variance model of measured data, C(t), was defined as:
 inter = 0.25 and slope = 0.1, an error process with a coefficient of variation of 10%.
In the second stage, the population pharmacokinetics were determined by using linear mixed-effects modeling as implemented in R (version 2.4.1; www.R-project.org):
 1 is the logarithm of the population mean CL/F; k are the coefficients for the effects of each covariate; and  and  describe the interpatient and intrapatient variability, respectively. (Both are assumed to have a mean of zero.)
Covariate Analysis Recursive partitioning (ie, classification and regression tree) also was used to test the interaction of covariates on dexamethasone CL/F. Because serum albumin concentration was a continuous variable, a cutoff value was chosen by using the model to split observations into two subgroups that best distinguished those with high versus low CL/F. At each step, the most predictive variable was determined by using linear mixed-effects modeling, and a cutoff for the predictive variable was chosen to split observations into two subgroups. This process was repeated for each subgroup until no variables were found to further accentuate the difference in CL/F between the groups.
Statistics
Observed Demographics We studied 212 courses on week 7 and 206 courses on week 8 in 214 patients (99 females and 115 males). Altogether, 407 courses were assessable, and 11 courses were excluded because the patients vomited within 1 hour of the dose, problems occurred with the IV line, or there was poor compliance in dosing or in blood sample collection. One hundred thirteen patients were assigned to the low-risk arm (median age, 4.17 years; range, 1.25 to 18.4 years), and 101 patients were assigned to the standard/high-risk arms combined (median age, 8.17 years; range, 1.00 to 18.8 years).
Pharmacokinetic Parameter Estimates
Glucocorticoids are substrates for and inducers of drug-metabolizing enzymes and transporters.14-17 Daily dexamethasone could cause autoinduction of clearance, inhibition of clearance, or a combination of both. To address intrapatient variability, we studied each patient twice, on day 1 (week 7) and day 8 (week 8) of an 8-day course of dexamethasone. The population mean CL/F at week 7 (15.5 L/h/m2; standard error, 0.84 L/h/m2) was greater than that at week 8 (12.4 L/h/m2; standard error, 0.54 L/h/m2; P = .004; Appendix Fig A1, online only). The mean ka was 1.5 hours–1; mean V/F was 46.8 L/m2; mean ke was 0.3 hours–1; and mean t1/2 was 2.3 hours (Appendix Table A2, online only).
Analysis of Covariates Albumin concentrations of patients on the standard/high-risk arms were less than those of patients on the low-risk arm at week 7 (P < .001); the means (± standard deviations) were 3.13 ± 0.63 g/dL and 4.14 ± 0.32 g/dL, respectively (Fig 2A). This finding is consistent with hypoalbuminemia caused by asparaginase, to which patients on the standard/high-risk arm were more exposed (Table 1). Moreover, there was a substantial decrease in the albumin concentration in patients in the low-risk arm from weeks 7 to 8 (P < .001), which reflected the reintroduction of asparaginase during week 7 to the low-risk arm (Fig 2A).
An allergy to asparaginase before reinduction was more common among patients in the standard/high-risk arms (31 of 101) than among those on the low-risk arm (1 of 113). In the standard/high-risk arms, serum albumin and dexamethasone CL/F were significantly greater in patients who had developed an allergy to asparaginase than in those who had not (P = .002 and .04, respectively; Fig 2B). Allergy may inactivate asparaginase, and this inactivation results in lower systemic exposure to asparaginase. The net result would be less asparaginase-mediated inhibition of protein synthesis and, thus, a greater serum albumin.18 We built multivariate models to assess how covariates of CL/F might interact when we included all patients (n = 355 courses in 207 patients) and when we excluded the outlying courses (n = 348 courses in 204 patients; Table 2). The treatment arm, use of ketoconazole or fentanyl, age, and serum albumin were associated with CL/F. The inter- and intrapatient variabilities (expressed as a CV %) for CL/F of the final model decreased to 40% and 45%, respectively, from 51% and 51%, respectively, for CL/F of the base model when the outliers were included and to 32% and 41%, respectively, from 44% and 46%, respectively, when the outliers were excluded.
Classification and regression tree analysis was also used to assess covariates for CL/F. Serum albumin was the most significant predictor; results were similar when the outlying values (n = 7) were excluded (Fig 3) or included (data not shown). The best cutoff value for distinguishing those with low CL/F from those with high CL/F was 3.35 g/dL. In patients with low serum albumin, only age (with a cutoff of 10 years) was a significant predictor of CL/F. In those with high serum albumin, the most important determinant of CL/F was treatment arm: patients in the standard/high-risk arms had a lower CL/F than those in the low-risk arm. In patients in the low-risk arm, CL/F was lower at week 8 than at week 7 (Fig 3).
Dexamethasone pharmacokinetic data in patients with cancer have been lacking. The present study is the first pharmacokinetic evaluation of dexamethasone in children with cancer or in any group of patients with ALL. The paucity of prior data is surprising, given the extent of dexamethasone use in oncology. Dexamethasone is a critical component of modern chemotherapy regimens for ALL, but there is uncertainty as to the optimal dosage; most trials use 6, 8, 10, or 12 mg/m2/d. We observed high inter- and intraindividual pharmacokinetic variabilities (CV of 46% and 53%, respectively) in dexamethasone CL/F, which resulted in a more than 10-fold variability in systemic exposure to the drug at a uniform dosage of 8 mg/m2/d; this variability dwarfs that anticipated when weighing, for example, 8 versus 12 mg/m2/d for administered doses. The extensive variability led to an examination of covariates for CL/F, especially of predictors likely to change within patients and thus to explain the high intraindividual variability. The multivariate analysis showed that serum albumin and the treatment arm were strongly associated with CL/F; both covariates were plausibly linked by the more intensive use of asparaginase in the standard/high-risk arms than in the low-risk arm. Univariate analyses of other covariates indicated additional variables that may impact dexamethasone pharmacokinetics in other clinical settings. Serum albumin was positively associated with CL/F (Fig 1B). A correlation between pharmacokinetic parameters and serum albumin has also been found for other drugs that are eliminated hepatically.19-21 Serum albumin is a measure of hepatic function22-24 and reflects the hepatic capacity to synthesize protein. Hypoalbuminemia is a well-recognized effect of asparaginase.25 Although hypoalbuminemia may be caused by other factors, we hypothesize that albumin was variable partly because of prior asparaginase use and that decreased albumin may be a biomarker of impaired hepatic synthesis of proteins involved in dexamethasone clearance.26 The albumin concentration on the standard/high-risk arms was approximately 25% lower than on the low-risk arm (Fig 2A), which was consistent with greater prior exposure to asparaginase in the former group (Table 1). The dexamethasone CL/F (Fig 2B) and the serum albumin changes (Fig 2A) as a result of asparaginase explain much of the inter- and intraindividual variabilities in CL/F. Low albumin predicted low clearance; however, among those with low albumin concentrations, age predicted clearance (Fig 3). In those with greater albumin, patients on the low-risk arm (who were younger and with less asparaginase) had a greater CL/F, and clearance at week 8 was less than that at week 7, possibly because of the asparaginase received during week 7 (Fig 3). The decrement in albumin from weeks 7 to 8 was much more impressive on the low-risk arm than on the standard/high-risk arms (Fig 2A). This result is consistent with the idea that the intensive asparaginase on the standard/high-risk arms had already caused hypoproteinemia by week 7; further treatment had minimal further effects on serum albumin, whereas patients on the low-risk arm were more susceptible to a decrement in albumin by the time they received asparaginase in week 7. Asparaginase plasma exposure is lowered by allergy, which is often accompanied by inactivating antibodies.18,27 At the time of dexamethasone pharmacokinetic analyses, allergy was more common on the standard/high-risk arms than on the low-risk arm. Serum albumin was greater in patients with clinical allergy by the start of the reinduction phase; this result is consistent with the hypothesis that systemic exposure to asparaginase was lower in patients with an asparaginase allergy (Fig 2B). Likewise, the dexamethasone CL/F was greater in those who had an asparaginase allergy (Fig 2B). Thus, patients with an allergy to asparaginase may be doubly disadvantaged in terms of an antileukemic effect—the allergy may be associated with the direct antibody-mediated inactivation of asparaginase, and the lower exposure to asparaginase may result in less hypoproteinemic effects and greater clearance of dexamethasone. Although dexamethasone is bound (approximately 80%) to plasma proteins, a finding of hypoalbuminemia in association with decreased CL/F is not consistent with a plasma protein-binding mechanism. Lower albumin concentrations would result in increased concentrations of unbound dexamethasone, but, because dexamethasone is a restrictively cleared drug, this increase in the unbound dexamethasone concentration should be accompanied by an increase in the clearance of the free drug, which would result in no net increase in exposure or total drug clearance.28 However, we observed increased clearance with an increased serum albumin concentration. Thus, at least in this setting, it appears that albumin's effect on CL/F is more likely to reflect hepatic clearance than plasma protein- binding. As is true for other agents,29 dexamethasone had a faster CL/F in younger than in older children. Although the effect of age on clearance was complicated by the fact that younger children were more likely to be treated on the low-risk arm rather than the standard/high-risk arms, the multivariate analyses (Table 2; Fig 3) suggest an independent effect of age. The dexamethasone CL/F in a patient at 19 years was 102.5% less than in a patient at 5 years (from 7.8 to 15.8 L/h/m2), which was consistent with greater toxicity among older children. Moreover, the dexamethasone clearance observed in adult studies (Table 3) was closer to that which we observed in older rather than younger children. Thus, adults are exposed to twice the active drug as children are when given the same dose.
The univariate analysis (Appendix Table A3, online only) of drug interactions may have been confounded by covariance with other characteristics (eg, serum albumin), as the dexamethasone CL/F paradoxically increased with the use of CYP3A4 substrates or inhibitors and decreased with the use of CYP3A4 inducers. In the multivariate analysis (Fig 3), however, these drugs had negligible effects on CL/F relative to more penetrant influences, such as albumin, age, week of therapy, and treatment arm (ie, prior asparaginase use). Plasma sampling was limited to five samples over 8 hours. For this reason, confidence intervals for pharmacokinetic parameters estimates were wide. Nevertheless, estimates of CL/F were comparable to or slightly greater than those reported in adults (Table 3). Likewise, our estimates of ka and V/F (Table 3) were also comparable to those found previously.10,30,32 In conclusion, dexamethasone pharmacokinetics displayed substantial inter- and intrapatient variability. Much of the variability was accounted for by variability in the serum albumin concentration, which in turn was affected by the intensity of prior asparaginase treatment. The dexamethasone CL/F was greater in younger than in older children, which resulted in almost twice the systemic exposure in adults than in younger children given the same dose. Our findings indicated that host- and treatment-related factors greatly affect systemic exposure to dexamethasone and could account for variable responses to this widely used agent.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a "U" are those for which no compensation was received; those relationships marked with a "C" were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Employment or Leadership Position: None Consultant or Advisory Role: Mary V Relling, OPI Pharmaceuticals (C) Stock Ownership: None Honoraria: Ching-Hon Pui, Enzon Corporation Research Funding: None Expert Testimony: None Other Remuneration: None
Conception and design: Mary V. Relling Financial support: Mary V. Relling Administrative support: Ching-Hon Pui, Mary V. Relling Provision of study materials or patients: Ching-Hon Pui, Mary V. Relling Collection and assembly of data: Lei Yang, Xiangjun Cai, Nancy Kornegay Data analysis and interpretation: Lei Yang, John C. Panetta, Wenjian Yang, Deqing Pei, Cheng Cheng Manuscript writing: Lei Yang, John C. Panetta, Mary V. Relling Final approval of manuscript: Lei Yang, John C. Panetta, Xiangjun Cai, Wenjian Yang, Deqing Pei, Cheng Cheng, Nancy Kornegay, Ching-Hon Pui, Mary V. Relling
We thank our protocol coinvestigators, clinical staff, research nurses, the patients, and their parents for their participation. We also thank David Tran and Nancy Duran for laboratory work, Donald Samulack for helping with scientific editing, and Mark Wilkinson for computing assistance.
Supported by Grants No. CA 51001 and CA 21765 from the National Institutes of Health, National Cancer Institute; by a Center of Excellence grant from the state of Tennessee; and by the American Lebanese Syrian Associated Charities. C.-H.P. is an American Cancer Society Professor. Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Smyth AC, Wiernik PH: Combination chemotherapy of adult acute lymphocytic leukemia. Clin Pharmacol Ther 19:240-245, 1976[Medline] 2. Smets LA, Salomons G, van den BJ: Glucocorticoid induced apoptosis in leukemia. Adv Exp Med Biol 457:607-614, 1999[Medline] 3. Fauci AS, Dale DC, Balow JE: Glucocorticosteroid therapy: Mechanisms of action and clinical considerations. Ann Intern Med 84:304-315, 1976 4. Cassileth PA, Lusk EJ, Torri S, et al: Antiemetic efficacy of dexamethasone therapy in patients receiving cancer chemotherapy. Arch Intern Med 143:1347-1349, 1983 5. Kris MG, Hesketh PJ, Somerfield MR, et al: American Society of Clinical Oncology guideline for antiemetics in oncology: Update 2006. J Clin Oncol 24:2932-2947, 2006 6. Cunningham D, Evans C, Gazet JC, et al: Comparison of antiemetic efficacy of domperidone, metoclopramide, and dexamethasone in patients receiving outpatient chemotherapy regimens. Br Med J (Clin Res Ed) 295:250, 1987 7. Mattano LA, Jr., Sather HN, Trigg ME, et al: Osteonecrosis as a complication of treating acute lymphoblastic leukemia in children: A report from the Children's Cancer Group. J Clin Oncol 18:3262-3272, 2000 8. Puisset F, Dalenc F, Chatelut E, et al: Dexamethasone as a probe for vinorelbine clearance. Br J Clin Pharmacol 60:45-53, 2005[CrossRef][Medline] 9. Puisset F, Chatelut E, Dalenc F, et al: Dexamethasone as a probe for docetaxel clearance. Cancer Chemother Pharmacol 54:265-272, 2004[Medline] 10. Richter O, Ern B, Reinhardt D, et al: Pharmacokinetics of dexamethasone in children. Pediatr Pharmacol 3:329-337, 1983 11. Pui CH, Relling MV, Sandlund JT, et al: Rationale and design of Total Therapy Study XV for newly diagnosed childhood acute lymphoblastic leukemia. Ann Hematol 83:S124-S126, 2004 (suppl 1)[CrossRef][Medline] 12. D’Argenio DZ, Schumitzky A: ADAPT II user's guide. Los Angeles, Biomedical Simulations Resource, USC, 1990 13. Steimer JL, Mallet A, Golmard JL, et al: Alternative approaches to estimation of population pharmacokinetic parameters: Comparison with the nonlinear mixed-effect model. Drug Metab Rev 15:265-292, 1984[Medline] 14. Xu C, Li CY, Kong AN: Induction of phase I, II and III drug metabolism/transport by xenobiotics. Arch Pharm Res 28:249-268, 2005[Medline] 15. Rushmore TH, Kong AN: Pharmacogenomics, regulation and signaling pathways of phase I and II drug metabolizing enzymes. Curr Drug Metab 3:481-490, 2002[CrossRef][Medline] 16. Demeule M, Jodoin J, Beaulieu E, et al: Dexamethasone modulation of multidrug transporters in normal tissues. FEBS Lett 442:208-214, 1999[CrossRef][Medline] 17. Jugert FK, Agarwal R, Kuhn A, et al: Multiple cytochrome P450 isozymes in murine skin: Induction of P450 1A, 2B, 2E, and 3A by dexamethasone. J Invest Dermatol 102:970-975, 1994[CrossRef][Medline] 18. Hak LJ, Relling MV, Cheng C, et al: Asparaginase pharmacodynamics differ by formulation among children with newly diagnosed acute lymphoblastic leukemia. Leukemia 18:1072-1077, 2004[CrossRef][Medline] 19. Twelves CJ, Dobbs NA, Gillies HC, et al: Doxorubicin pharmacokinetics: The effect of abnormal liver biochemistry tests. Cancer Chemother Pharmacol 42:229-234, 1998[CrossRef][Medline] 20. Kovarik JM, Sabia HD, Figueiredo J, et al: Influence of hepatic impairment on everolimus pharmacokinetics: Implications for dose adjustment. Clin Pharmacol Ther 70:425-430, 2001[Medline] 21. Kovarik JM, Schmouder RL, Hartmann S, et al: Fingolimod (FTY720) in severe hepatic impairment: Pharmacokinetics and relationship to markers of liver function. J Clin Pharmacol 46:149-156, 2006 22. Limdi JK, Hyde GM: Evaluation of abnormal liver function tests. Postgrad Med J 79:307-312, 2003 23. Burke MD: Hepatic function testing. Postgrad Med 64:177- 182, 1978[Medline] 24. Mahl TC: Approach to the patient with abnormal liver tests. Lippincotts Prim Care Pract 2:379-389, 1998[Medline] 25. Petros WP, Rodman JH, Relling MV, et al: Variability in teniposide plasma protein binding is correlated with serum albumin concentrations. Pharmacotherapy 12:273-277, 1992[Medline] 26. Burton ME, Show LM, Schentag JJ, et al: Drug metabolism, transport, and the influence of hepatic disease, in Troy D (ed): Applied Pharmacokinetics: Principles of Therapeutic Drug Monitoring. Baltimore, MD, Lippincott Williams & Wilkins, 2005, pp 135-142 27. Woo MH, Hak LJ, Storm MC, et al: Anti-asparaginase antibodies following E. coli asparaginase therapy in pediatric acute lymphoblastic leukemia. Leukemia 12:1527-1533, 1998[CrossRef][Medline] 28. Wilkinson GR, Shand DG: Commentary: A physiological approach to hepatic drug clearance. Clin Pharmacol Ther 18:377-390, 1975[Medline] 29. Crom WR, Glynn-Barnhart AM, Rodman JH, et al: Pharmacokinetics of anticancer drugs in children. Clin Pharmacokinet 12:168-213, 1987[Medline] 30. Rose JQ, Yurchak AM, Meikle AW, et l: Effect of smoking on prednisone, prednisolone, and dexamethasone pharmacokinetics. J Pharmacokinet Biopharm 9:1-14, 1981[CrossRef][Medline] 31. O'Sullivan BT, Cutler DJ, Hunt GE, et al: Pharmacokinetics of dexamethasone and its relationship to dexamethasone suppression test outcome in depressed patients and healthy control subjects. Biol Psychiatry 41:574-584, 1997[CrossRef][Medline] 32. Brophy TR, McCafferty J, Tyrer JH, et al: Bioavailability of oral dexamethasone during high dose steroid therapy in neurological patients. Eur J Clin Pharmacol 24:103-108, 1983[CrossRef][Medline] Submitted August 2, 2007; accepted November 20, 2007.
Related Editorial
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2008 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
8 hours). 
