Drug-induced hepatotoxicity is a well-recognized adverse. Incidence of Statin Hepatotoxicity in Patients With Hepatitis C

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2006;4: Incidence of Statin Hepatotoxicity in Patients With Hepatitis C SHIRIN KHORASHADI,* NOELLE K. HASSON,* and RAMSEY C. CHEUNG, *Pharmacy Service, and Division of Gastroenterology and Hepatology, VA Palo Alto Health Care System, Palo Alto, California; and Stanford University, Stanford, California See CME exam on page 806; see editorial on page 838. Background & Aims: Statins are considered contraindicated in patients with chronic liver disease. Our objective was to determine the risk of developing hepatotoxicity from statin therapy in hyperlipidemic patients with hepatitis C. Methods: Changes in liver biochemistry values within 12 months compared with baseline were determined in 3 cohorts matched for age, sex, and body mass index: (I) 166 anti hepatitis C virus (HCV)-positive hyperlipidemic veterans who were initiated on statin therapy; (II) 332 anti HCV-positive veterans who had not received statin therapy; and (III) 332 anti HCV-negative hyperlipidemic veterans who were initiated on statin therapy. An increase in liver biochemistry values was defined as mild-moderate or severe as proposed in a previous study on statin hepatotoxicity in a non hepatitis C population. Results: In patients with hepatitis C, statin therapy (cohort I) was associated with a higher incidence of mild-moderate liver biochemistry value increases compared with those not on statin therapy (cohort II) (22.9% vs 13.3%, respectively, P.009), but a lower incidence of severe increases (1.2% vs 6.6%, respectively, P.015). Among patients started on statin therapy (cohorts I and III), the incidence of mild-moderate liver biochemistry value increases (22.9% vs 16.3%, respectively, P.094), severe increases (1.2% vs 1%, respectively, P.874), or discontinuation of statin therapy as a result of hepatotoxicity (21.6% vs 9.2%, respectively, P.147) were similar in hepatitis C positive and hepatitis C negative patients. Conclusions: Statin therapy was not associated with a higher risk of severe hepatotoxicity in patients with chronic hepatitis C and appeared safe. Drug-induced hepatotoxicity is a well-recognized adverse reaction, and the proportion of cases of hepatic disease known to be drug related continues to increase. 1 3 In the United States, these occurrences account for more than 50% of the cases of acute liver failure. 4,5 The mechanism of action of drug-induced hepatic disease is either a result of the drug itself or a result of reactive metabolites of the drug that covalently bind to hepatocytes and result in either idiosyncratic (unpredictable), intrinsic (predictable), or immunoallergic hepatitis. 1,5,6 Most hepatotoxic drug reactions are idiosyncratic and are neither dose-dependent nor predictable. 5 Recently, genetic variations in systems of biotransformation or detoxification also were reported to contribute to drug hepatotoxicity. 6 One of the most commonly prescribed class of medications is statins, the drugs of choice for low-density lipoprotein lowering treatment in patients at risk for cardiovascular disease. 7 Although statins generally are well tolerated, they have been reported to lead to increases in transaminase level (.2% 2.7% of patients) and, rarely, symptomatic liver disease including fatal acute liver failure. 8,9 Liver enzyme level increases typically occur within the first 3 12 months after initiation of therapy, are asymptomatic, and are more prevalent with higher doses. Because of these reports, experts and manufacturers recommend routine monitoring of liver function test (LFT) values in patients receiving chronic statin therapy, and advise against statin therapy in patients with active or chronic liver disease. 7,10,11 These recommendations have been challenged recently because of a paucity of data supporting an increased risk of liver enzyme level abnormalities in the low-risk general population. In a recent meta-analysis, De Denus et al 8 found that pravastatin, lovastatin, and simvastatin at low-tomoderate doses were not associated with an increased risk of liver enzyme level abnormalities compared with placebo. Because of these findings, the latest clinical guidelines on the management of dyslipidemia in patients with type 2 diabetes mellitus did not recommend routine monitoring of liver function unless baseline LFTs were abnormal. 12 The management is less clear for those with abnormal baseline liver enzyme levels or chronic liver diseases. When compared with controls with normal baseline liver Abbreviations used in this paper: ALT, alanine transaminase; anti- HCV, antibody to hepatitis C virus; AST, aspartate transaminase; HCV, hepatitis C virus; LFT, liver function test; VA, Veterans Affairs by the American Gastroenterological Association Institute /06/$32.00 doi: /j.cgh

2 July 2006 STATIN HEPATOTOXICITY IN HEPATITIS C 903 enzyme levels, patients with presumed nonalcoholic fatty liver disease and increased baseline liver enzyme levels did not have a higher incidence of severe increases in liver biochemistry values within the first 6 months after initiation of a statin. 13,14 However, the study excluded patients with chronic hepatitis C in whom the risk of hepatotoxicity largely is unknown. Currently in clinical practice, primary care physicians often are reluctant to prescribe statins to treat hyperlipidemic patients with chronic hepatitis C owing to concerns of hepatotoxicity. The objective of this study was to determine in patients with hepatitis C whether statin therapy increases the risk for developing hepatotoxicity when compared with hepatitis C negative patients on statin therapy and with hepatitis C positive patients not on statin therapy. Materials and Methods Study Design This study was approved by the Institutional Review Board of Stanford University. All patients were veterans receiving care from the Veterans Affairs (VA) Palo Alto Health Care System, which is a multicampus tertiary referral teaching hospital affiliated with Stanford University. The VA Palo Alto Health Care System consists of 3 major divisions and multiple outpatient clinics in northern California. The hospital uses an elaborate electronic medical record system that includes clinic visit notes, diagnostic codes of the clinic encounter, all prescriptions, and laboratory data in a searchable form. Potential study patients were identified initially using the VA pharmacy and laboratory database. Three cohorts of patients, 1 case study group (cohort I) and 2 control groups (cohorts II and III), were identified retrospectively. Cohort I included all patients who tested positive for antibody to hepatitis C virus (anti-hcv) by third generation enzyme immunoassay (EIA III; Abbott Laboratories, Abbott Park, IL) between October 1, 2002 and September 30, 2003 and were started on a statin after the diagnosis of HCV. Cohort II consisted of a matched sample of anti-hcv positive patients tested between October 1, 2002 and September 30, 2003 who had not received statin therapy. Cohort III included a matched sample of anti-hcv negative patients who started statin therapy between October 1, 2002 and September 30, Patients from cohorts II and III were matched to those in cohort I based on age, sex, and body mass index. In cohorts I and III, patients who did not have LFTs within 1 year before and after initiation of the statin therapy were excluded. Similarly, patients in cohort II who did not have LFTs within 1 year before and after HCV diagnosis were excluded. Data Collection By using the hospital computer database, pertinent demographic data such as age, sex, height, weight, and ethnicity were collected. In addition, data on laboratory values (eg, LFTs and lipid panel), name of statin prescribed, and medication fill dates also were collected. Information regarding an active alcohol problem was obtained from patients administrative files. An active alcohol problem was defined in this study as the inclusion of an alcohol-related diagnosis (International Classification of Diseases Version 9 codes 571.1, 303, , , 303.9, , , 305, , and ) in any clinic or hospital encounter during the study period. Concomitant use of potentially hepatotoxic medications within the study period was determined from pharmacy and medical records. The LFTs consisted of aspartate transaminase (AST), alanine transaminase (ALT), and serum bilirubin values. For the purpose of this study, baseline values were defined as LFTs obtained within 1 year before the initiation of a statin (cohorts I and III) or the date of hepatitis C diagnosis (cohort II). The LFTs obtained within 1 year after the initiation of the statin in cohorts I and III were used to determine the hepatotoxic effect of the statin. In cohort II, LFTs within 1 year after HCV diagnosis were used to determine the fluctuation of liver enzyme levels over time. Statin therapy was considered as discontinued when the medication status was changed from active to discontinued in the prescription database. This was confirmed further by reviewing physician progress notes to include only those in whom the statin was discontinued owing to LFT increases. Definition of Hepatotoxicity The normal ranges in our laboratory were 0 45 U/L for ALT and 0 41 U/L for AST. Values above the upper limit of normal (ie, ALT 45 U/L, AST 41 U/L) were considered abnormal. For ease of comparison with published data, hepatotoxicity was defined in the same way as proposed by Chalasani et al. 13 The severity of increase was defined according to the degree of abnormality of the follow-up LFTs as compared with the baseline values. Mild-moderate increases were defined as an AST or ALT increase up to 10 times the upper limit of normal (for those with normal baseline values), or from the baseline level if patients had increased transaminase values at study entry. Severe increases were defined as serum bilirubin value greater than 3 mg/dl (regardless of AST or ALT values) or an increase in the AST or ALT value of more than 10 times the upper limit of normal (or of baseline level if patients had increased baseline transaminase values). For patients who had multiple LFTs performed within the study period, the highest value was used for comparison. Statistical Analysis For the data that was not distributed normally, results are presented as medians and 25th to 75th percentiles. Nonparametric statistic tests were used where applicable. Patient demographic data were analyzed using the Mann Whitney rank-sum test for interval data and the 2 test with Yates correction for nominal data comparisons between the case and each control group. To compare the change in liver enzyme values between the groups, the Mann Whitney rank-sum test was used. The 2 test with Yates correction was used to compare the percentage of patients with increases in liver

3 904 KHORASHADI ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 7 Table 1. Baseline Demographics HCV positive no statin (cohort II) (n 332) HCV positive statin (cohort I) (n 166) HCV negative statin (cohort III) (n 332) Median age, y (25th, 75th percentiles) 54 (50, 60) 53 (50, 56) 55 (52, 61) Sex, % men Median body mass index, kg/m 2 29 (26, 32) 28 (26, 31) 30 (27, 32) (25th, 75th percentiles) Ethnicity, n (%) American Indian 8 (2.4) 0 0 Asian 3 (.9) 1 (.6) 5 (1.5) African American 33 (9.9) 14 (8.4) 27 (8.1) Caucasian 144 (43.4) 76 (45.8) 114 (34.3) Hispanic 33 (9.9) 14 (8.4) 11 (3.3) Unknown 111 (33.4) 61 (36.7) 175 (52.7) enzyme values. Statistical tests were performed using Sigma Stat (version 2; Jandel, Sausalito, CA). Results A total of 830 patients were included in the study: 166 patients in cohort I and 332 matched patients each in cohorts II and III. The baseline characteristics were similar in all 3 groups (Table 1). The majority of patients (98%) were men, which is consistent with the VA patient population. The baseline laboratory values are summarized in Table 2. In patients with hepatitis C, the baseline median liver biochemistry values were similar in patients who received a statin therapy (cohort I) compared with those who did not receive a statin (cohort II). However, when comparing the 2 cohorts of patients on statin therapy (cohorts I and III), patients with hepatitis C (cohort I) had higher baseline median liver enzyme values compared with the hepatitis C negative individuals (cohort III) (P.001). Baseline median values of the lipid panels in patients on statin therapy were similar between hepatitis C positive and hepatitis C negative groups, whereas those not on a statin therapy had significantly lower baseline lipid panel values. As shown in Figure 1, the median percentage change of liver enzyme levels from baseline was small in all 3 cohorts. Patients with hepatitis C on statin therapy (cohort I) had a lower percentage change in median liver enzyme levels from baseline compared with those not on statin therapy (cohort II) (AST level, 1% vs 5%, respectively, P.032; ALT level, 7.3% vs 6.9%, respectively, P.001) and compared with the hepatitis C negative group on statin therapy (cohort III) (AST level, 1% vs 5%, respectively, P.004; ALT level, 7.3% vs 4.8%, respectively, P.002). The overall percentage of patients with any liver biochemistry value increase was similar in all 3 cohorts (Figure 2). A higher percentage of patients with hepatitis C who started statin therapy developed mild-moderate increases compared with those not on statin therapy (22.9% vs 13.3%, respectively, P.009). However, this group had a lower percentage of patients with severe increases compared with those hepatitis C patients not on statin therapy (1.2% vs 6.6%, respectively, P.015). Among all the patients who were put on statin therapy (cohort I vs III), the percentage of patients with hepatitis C who developed mild-moderate increases (22.9% vs Table 2. Baseline Laboratory Values HCV positive no statin (cohort II) (n 332) HCV positive statin (cohort I) (n 166) HCV negative statin (cohort III) (n 332) Median LFT, U/L (25th, 75th percentile) a AST 32 (24, 51) 36 (26, 48) 26 b (21, 32) ALT 39 (24, 67) 40 (26, 66) 27 b (20, 40) Median, mg/dl (25th, 75th percentile) Cholesterol 157 b (140, 196) 205 (177, 238) 215 (185, 242) LDL 96 b (75, 126) 128 (105, 156) 133 (112, 161) HDL 42 (33, 49) 41 (35, 51) 41 (35, 47) TG 109 b (75, 165) 132 (101, 185) 155 (102, 224) LDL, low-density lipoprotein; HDL, high-density lipoprotein; TG, triglyceride. a The mean number of LFT for cohorts II, I, III was 2.4, 2.8, and 2.1 at baseline and 3.6, 3.2, and 3.0 during the follow-up period, respectively. b P.001 vs cohort I.

4 July 2006 STATIN HEPATOTOXICITY IN HEPATITIS C 905 Figure 1. Percent change in liver enzyme levels from baseline to within 12 months after initiation of a statin therapy or diagnosis of hepatitis C in the no statin group. *P.05 compared with the HCV-positive statin group., HCV positive and no statin (n 332);, HCV positive with statin (n 166);, HCV negative with statin (n 332). 16.3%, respectively, P.094) or severe increases (1.2% vs 1%, respectively, P.874) were similar to the anti HCV-negative group. There was a trend but no significant difference in the percentage of patients with liver biochemistry value increases who discontinued statin therapy between cohorts I and III (21.6% vs 9.2%, respectively, P.147). The most commonly prescribed statins in both cohorts (I and III) were simvastatin (53% and 52%, respectively) and lovastatin (28% and 25%, respectively). We also wanted to address other potential cofounding factors such as concomitant medication and alcohol use. The percentage of patients on statin therapy with an active alcohol problem were not different between cohorts I and III (12.7% vs 9.9%, P.445). However, in patients with hepatitis C, a higher percentage of patients not on a statin therapy had evidence of an alcohol problem compared with those on a statin therapy (24.1% vs 12.7%, P.004). In patients with hepatitis C (cohorts I and II), the percentage of patients who received hepatitis C treatment during the study period were similar (3% vs 2.4%). The most frequently prescribed concomitant medications with potential hepatotoxicity during the study period in the 3 cohorts were acetaminophen, ranitidine, bupropion, valproic acid, rosiglitazone, nefazadone, amoxicillin/clavulanate, phenytoin, olanzapine, and quetiapine. There was no significant difference with the use of these medications between the 3 groups (data not shown). Discussion Recent evidence suggests the incidence of hepatotoxicity from statin therapy is low. The mechanism of increase of liver enzyme level, usually transient, is unknown but might be related to the lipid-lowering effect of the medication itself. 15 The current prescribing practice of monitoring LFTs and the labeling recommendation to avoid statins in patients with chronic liver disease have been challenged recently in patients with type 2 diabetes mellitus 12 and nonalcoholic fatty liver disease, 13,14 respectively. Unfortunately, there is a paucity of information on patients with chronic hepatitis C, which is the most common form of viral hepatitis. The third National Health and Nutrition Examination Survey estimated that approximately 3.9 million people (1.8% of the population) are infected with hepatitis C in the United States, and the prevalence was even higher among the age groups with the highest incidence of hyperlipidemia. 16 The prevalence of hepatitis C is much higher among veterans, and hyperlipidemia is especially common because most veterans are in their 50s. Certainly many providers feel uncomfortable prescribing statin therapy in hyperlipidemic patients with chronic hepatitis C owing to uncertain risks, despite the potential benefits. 7,10 In a post hoc subgroup analysis, Chalasani et al 13 found in a group of 1111 individuals with detectable hepatitis C antibody who were not on statin therapy (similar to cohort II of the current study) that there was a significantly higher frequency of mild-moderate or severe increases in liver biochemistry values compared Figure 2. Frequency and degree of hepatotoxicity within 12 months after initiation of a statin therapy or diagnosis of hepatitis C in the no statin group. Severity of increases in liver biochemistry values as defined in text. *P.05 compared with the HCV-positive statin group., HCV positive and no statin (n 332);, HCV positive with statin (n 166);, HCV negative with statin (n 332).

5 906 KHORASHADI ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 7 with non hepatitis C patients with increased baseline liver enzyme levels who were put on statins. This observation supported the notion that patients with hepatitis C are more prone to liver enzyme level abnormalities that were not related to drug therapy, and this was confirmed by the current study. In the current study, changes from baseline liver enzyme levels in the 3 cohorts were statistically significant between certain cohorts, but they were not clinically relevant because the highest change in median level was less than 8%. A more clinically relevant end point was the percentage of patients with severe increases and the percentage of patients in whom the statin was discontinued by the physician owing to LFT increase. In patients with hepatitis C, statin therapy was associated with the highest risk for developing mild-moderate increases in liver biochemistry values; interestingly, those not on a statin therapy were at higher risk for developing severe increases. Furthermore, statin therapy did not increase the risk of life-threatening severe increases in liver biochemistry values in patients with hepatitis C compared with the hepatitis C negative controls. Hence, other concomitant factors such as alcohol use also may play a role in causing liver enzyme level abnormalities in hepatitis C patients not on a statin. Indeed, this group had a higher percentage of patients with an active alcohol problem compared with those on statin therapy. It also was possible that only hepatitis C patients with stable LFTs over a longer period of time and not using alcohol, and those perceived to have mild liver disease, were started on statin therapy by the physicians. Even though we did not find a difference in the percentage of patients who underwent liver biopsy examination and the mean fibrosis stage on liver biopsy examination between the HCV-positive veterans on statin and those not on statin, a higher percentage in the latter group carried a diagnosis of cirrhosis (data not shown). The current study found that among the chronic hepatitis C patients not on statin therapy, 13.3% and 6.6% developed mild-moderate and severe increases in LFTs, respectively, which indicated the importance of including this group for comparison. The incidence of increases in liver biochemistry values in this untreated hepatitis C group was almost identical to that found by Chalasani et al. 13 However, we found a much higher incidence of mild-moderate increases in liver biochemistry values in our anti HCV-negative patients on statin compared with their control group with normal baseline LFTs on statin therapy. This might be related to the fact that we examined the liver biochemistry values over a 12-month instead of a 6-month period. In the HCV-negative group on statins, we also did not exclude those patients with abnormal baseline liver test results. The current study found that patients with hepatitis C who received statin therapy had an improvement in the median liver enzyme level from baseline compared with those who did not receive a statin therapy. This might not be relevant clinically other than it confirms the safety of statin therapy in these patients, but also supports the notion that treatment of hyperlipidemia actually may improve liver enzyme level abnormalities caused by fatty liver. 20 In the statin therapy group, anti HCV-negative patients had significantly lower baseline mean liver enzyme levels than those with hepatitis C (P.001), but the incidence of mild-moderate or severe increases was not significantly different between the 2 groups. Hence, our study showed that statin therapy does not increase the risk for developing hepatotoxicity in patients with hepatitis C compared with hepatitis C negative individuals. In this study, a higher percentage of anti HCVpositive patients discontinued statin therapy than the anti HCV-negative group, but the difference did not achieve statistical significance. The 9.2% discontinuation rate of statin therapy in the anti HCV-negative group was similar to a previously reported rate of 10.7% in patients with normal baseline liver enzyme levels. 13 The higher discontinuation rate in the anti-hcv group might be a reflection of the comfort level of the prescribing physicians because the incidence of severe increases was similar in both groups. It is likely that any increase in liver enzyme levels, even mild, will result in discontinuation of statins by the providers in patients who were hepatitis C positive, but less so in the hepatitis C negative group. We indeed found a much higher percentage of statin discontinuation was a result of LFT increases in the anti HCV-positive than the anti HCV-negative group (data not shown). The current study design is very similar to a previous study that excluded patients with chronic hepatitis C or evidence of alcohol abuse. 13 Major differences in the current study include a 12-month follow-up observation period instead of a 6-month period after initiation of statin therapy. In addition, cohorts were matched for body mass index in the current study because nonalcoholic fatty liver disease, commonly associated with central obesity, could be a cofounding factor in these hyperlipidemic patients regardless of their hepatitis C status. There also were limitations to this study. First, our patients predominantly were men. This study is particularly important for the VA, which has a much higher prevalence of HCV than the general population Second, only anti-hcv was used as a surrogate marker

6 July 2006 STATIN HEPATOTOXICITY IN HEPATITIS C 907 for HCV infection. Most patients who were tested for anti-hcv had 1 or more risk factors for HCV infection, and previous studies have shown that only 3.6% of veterans testing positive for anti-hcv represent false positives. 18 Our previous study found that 96% of 215 consecutive anti HCV-positive veterans were viremic. 19 Third, we assumed that patients received all their medication through the VA and were adherent with their statin therapy. Our experience was that the overwhelming majority of veterans who received care from the VA received all their prescriptions from the VA because of the excellent drug coverage benefit (unpublished observation). Patients must be taking their medication with at least some regularity according to the medication refill record. Finally, because of the extremely low incidence of idiosyncratic drug reactions, this study was not powered to detect any such reaction. In summary, the current study found that the incidence of mild-moderate or severe increases in liver biochemistry values and discontinuation of statins from hepatotoxicity were not significantly different between hepatitis C positive and hepatitis C negative groups. Among hepatitis C patients, those on statin therapy had a higher incidence of mild-moderate increases but a lower incidence of severe increases compared with those not on a statin. Our findings suggest that statins can be used safely in hyperlipidemic patients with chronic hepatitis C. References 1. Zimmerman HJ. Hepatology: a century of progress. Clin Liver Dis 2000;4: Pishvaian AC, Trope BW, Lewis JH. Drug-induced liver disease in Curr Opin Gastroenterol 2004;20: Novak D, Lewis JH. Drug-induced liver disease. Curr Opin Gastroenterol 2003;19: Lee WM. Acute liver failure in the United States. Semin Liver Dis 2003;22: Lee WM. Drug-induced hepatotoxicity. N Engl J Med 2003;349: Larrey D, Pageaux GP. Genetic predisposition to drug-induced hepatotoxicity. J Hepatol 1997;26: Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 2001;285: De Denus S, Spinler SA, Miller K, et al. Statins and liver toxicity: a meta-analysis. Pharmacotherapy 2004;24: Russo MW, Galanko JA, Shresrha R, et al. Liver transplantation from drug induced liver injury in the United States. Liver Transpl 2004;10: Nash DT. A statin for a patient with hyperlipidemia and hepatitis C? Postgrad Med 2004;115: Pasternak RC, Smith SC Jr, Bairey-Merz CN, et al. ACC/AHA/ NHLBI Advisory on the use and safety of statins. J Am Coll Cardiol 2002;40: Snow V, Aronson MD, Hornbake ER, et al. Clinical Efficacy Assessment Subcommittee of the American College of Physicians. Lipid control in the management of type 2 diabetes mellitus: a clinical practice guideline from the American College of Physicians. Ann Intern Med 2004;140: Chalasani N, Aljadhey H, Kesterson J, et al. Patients with elevated liver enzymes are not at higher risk for statin hepatotoxicity. Gastroenterology 2004;126: Chalasani N. Statins and hepatotoxicity: focus on patients with fatty liver. Hepatology 2005;41: Tolman KG. Defining patient risks from expanded preventive therapies. Am J Cardiol 2000;85:15E 19E. 16. Alter MJ, Kruszon-Moran D, Nainan OV, et al. The prevalence of hepatitis C virus infection in the United States, 1988 through N Engl J Med 1999;341: Roselle GA, Danko LH, Kralovic SM, et al. National Hepatitis C Surveillance Day in the Veterans Health Administration of the Department of Veterans Affairs. Mil Med 2002;167: Briggs ME, Baker C, Hall R, et al. Prevalence and risk factors for hepatitis C virus infection at an urban Veterans Administration medical center. Hepatology 2001;34: Cheung RC. Epidemiology of hepatitis C virus infection in American veterans. Am J Gastroenterol 2000;95: Angulo P. Nonalcoholic fatty liver disease. N Engl J Med 2002; 346: Address requests for reprints to: Ramsey C. Cheung, MD, VA Palo Alto Health Care System (154C), 3801 Miranda Avenue, Palo Alto, California rcheung@stanford.edu; fax: (650) The authors would like to thank Sylvia Luong, PharmD, for proofreading the manuscript.