Advertisement

Oral Azole Antifungal Medications and Risk of Acute Liver Injury, Overall and by Chronic Liver Disease Status

  • Vincent Lo Re III
    Correspondence
    Corresponding author is Vincent Lo Re III, MD, MSCE, Center for Clinical Epidemiology and Biostatistics, 836 Blockley Hall, 423 Guardian Drive, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104-6021.
    Affiliations
    Division of Infectious Diseases, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Center for Pharmacoepidemiology Research and Training, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    Search for articles by this author
  • Dena M. Carbonari
    Affiliations
    Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Center for Pharmacoepidemiology Research and Training, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    Search for articles by this author
  • James D. Lewis
    Affiliations
    Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Center for Pharmacoepidemiology Research and Training, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Division of Gastroenterology, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    Search for articles by this author
  • Kimberly A. Forde
    Affiliations
    Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Center for Pharmacoepidemiology Research and Training, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Division of Gastroenterology, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    Search for articles by this author
  • David S. Goldberg
    Affiliations
    Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Center for Pharmacoepidemiology Research and Training, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    Search for articles by this author
  • K. Rajender Reddy
    Affiliations
    Center for Pharmacoepidemiology Research and Training, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Division of Gastroenterology, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    Search for articles by this author
  • Kevin Haynes
    Affiliations
    Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Center for Pharmacoepidemiology Research and Training, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    Search for articles by this author
  • Jason A. Roy
    Affiliations
    Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Center for Pharmacoepidemiology Research and Training, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    Search for articles by this author
  • Daohang Sha
    Affiliations
    Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    Search for articles by this author
  • Amy R. Marks
    Affiliations
    Division of Research, Kaiser Permanente Northern California, Oakland
    Search for articles by this author
  • Jennifer L. Schneider
    Affiliations
    Division of Research, Kaiser Permanente Northern California, Oakland
    Search for articles by this author
  • Brian L. Strom
    Affiliations
    Center for Clinical Epidemiology and Biostatistics, Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Center for Pharmacoepidemiology Research and Training, Perelman School of Medicine, University of Pennsylvania, Philadelphia

    Rutgers Biomedical and Health Sciences, Rutgers, the State University of New Jersey, Newark
    Search for articles by this author
  • Douglas A. Corley
    Affiliations
    Division of Research, Kaiser Permanente Northern California, Oakland
    Search for articles by this author
Published:November 17, 2015DOI:https://doi.org/10.1016/j.amjmed.2015.10.029

      Abstract

      Background

      Reports on associations between azole antifungal medications and acute liver injury are inconsistent and have not been based on liver-related laboratory tests. We evaluated incidence rates of acute liver injury associated with oral azole antifungals.

      Methods

      We conducted a cohort study among Kaiser Permanente Northern California members who initiated an oral azole antifungal in an outpatient setting during 2004-2010. We determined development of: (1) liver aminotransferases >200 U/L, (2) severe acute liver injury (coagulopathy with hyperbilirubinemia), and (3) acute liver failure. We calculated incidence rates of endpoints. Cox regression was used to determine whether chronic liver disease was a risk factor for outcomes.

      Results

      Among 195,334 azole initiators (178,879 fluconazole; 14,296 ketoconazole; 1653 itraconazole; 478 voriconazole; 28 posaconazole), incidence rates (events/1000 person-years [95% confidence intervals (CIs)]) of liver aminotransferases >200 U/L were similarly low with fluconazole (13.0 [11.4-14.6]), ketoconazole (19.3 [13.8-26.3]), and itraconazole (24.5 [10.6-48.2]). Rates were higher with voriconazole (181.9 [112.6-278.0]) and posaconazole (191.1 [23.1-690.4]), but comparable. Severe acute liver injury was uncommon with fluconazole (2.0 [1.4-2.7]), ketoconazole (2.9 [1.1-6.3]), and itraconazole (0.0 [0.0-11.2]), but more frequent with voriconazole (16.7 [2.0-60.2]) and posaconazole (93.4 [2.4-520.6]). One patient developed acute liver failure due to ketoconazole. Pre-existing chronic liver disease increased risks of aminotransferases >200 U/L (hazard ratio 4.68 [95% CI, 3.68-5.94]) and severe acute liver injury (hazard ratio 5.62 [95% CI, 2.56-12.35]).

      Conclusions

      Rates of acute liver injury were similarly low for fluconazole, ketoconazole, and itraconazole. Events were more common among voriconazole and posaconazole users but were comparable. Pre-existing chronic liver disease increased risk of azole-induced liver injury.

      Keywords

      Clinical Significance
      • Reports on associations between oral azole antifungal medications and acute liver injury are inconsistent and primarily based on spontaneous adverse event reports and prior analyses of acute liver injury diagnosis codes.
      • Absolute risks and rates of acute liver injury were low for fluconazole, ketoconazole, and itraconazole and were more common among users of voriconazole and posaconazole.
      • Pre-existing chronic liver disease was a strong risk factor for development of acute liver injury among azole users.
      Azole antifungal medications are prescribed as treatment for dermatophyte, mucocutaneous, or systemic fungal infections.
      • Como J.A.
      • Dismukes W.E.
      Oral azole drugs as systemic antifungal therapy.
      Ketoconazole may be prescribed as treatment of hormone-refractory prostate cancer.
      • Kim W.
      • Ryan C.J.
      Androgen receptor directed therapies in castration-resistant metastatic prostate cancer.
      Each azole antifungal has the potential to induce acute liver injury, characterized by elevations in liver aminotransferase levels or, in more serious cases, by hepatic dysfunction (severe acute liver injury, manifested by coagulopathy and hyperbilirubinemia
      • Lo Re 3rd, V.
      • Haynes K.
      • Goldberg D.
      • et al.
      Validity of diagnostic codes to identify cases of severe acute liver injury in the US Food and Drug Administration's Mini-Sentinel Distributed Database.
      ) or acute liver failure, defined by coagulopathy and hepatic encephalopathy.
      • Chalasani N.
      • Fontana R.J.
      • Bonkovsky H.L.
      • et al.
      Causes, clinical features, and outcomes from a prospective study of drug-induced liver injury in the United States.
      Recently, the US Food and Drug Administration (FDA) and European Medicines Agency (EMA) concluded that the risk of acute liver injury was higher for ketoconazole than other azole antifungals.

      European Medicines Agency. European Medicines Agency recommends suspension of marketing authorisations for oral ketoconazole. Benefit of oral ketoconazole does not outweigh risk of liver injury in fungal infections. EMA/458028/2013: 26 July 2013. Available at: www.ema.europa.eu/docs/en_GB/document_library/Press_release/2013/07/WC500146613.pdf. Accessed May 12, 2015.

      US Food and Drug Administration. FDA Drug Safety Communication: FDA limits usage of Nizoral (ketoconazole) oral tablets due to potentially fatal liver injury and risk of drug interactions and adrenal gland problems. UCM 36244: 26 July 2013. Available at: www.fda.gov/downloads/Drugs/DrugSafety/UCM362444.pdf. Accessed May 12, 2015.

      The EMA recommended that the marketing authorizations of oral ketoconazole-containing medications should be suspended throughout the European Union,

      European Medicines Agency. European Medicines Agency recommends suspension of marketing authorisations for oral ketoconazole. Benefit of oral ketoconazole does not outweigh risk of liver injury in fungal infections. EMA/458028/2013: 26 July 2013. Available at: www.ema.europa.eu/docs/en_GB/document_library/Press_release/2013/07/WC500146613.pdf. Accessed May 12, 2015.

      and the FDA issued a Drug Safety Communication (“black box warning”) recommending against ketoconazole's use, particularly in patients with chronic liver disease.

      US Food and Drug Administration. FDA Drug Safety Communication: FDA limits usage of Nizoral (ketoconazole) oral tablets due to potentially fatal liver injury and risk of drug interactions and adrenal gland problems. UCM 36244: 26 July 2013. Available at: www.fda.gov/downloads/Drugs/DrugSafety/UCM362444.pdf. Accessed May 12, 2015.

      However, these recommendations were based primarily on analyses of spontaneous adverse event reports.

      European Medicines Agency. European Medicines Agency recommends suspension of marketing authorisations for oral ketoconazole. Benefit of oral ketoconazole does not outweigh risk of liver injury in fungal infections. EMA/458028/2013: 26 July 2013. Available at: www.ema.europa.eu/docs/en_GB/document_library/Press_release/2013/07/WC500146613.pdf. Accessed May 12, 2015.

      US Food and Drug Administration. FDA Drug Safety Communication: FDA limits usage of Nizoral (ketoconazole) oral tablets due to potentially fatal liver injury and risk of drug interactions and adrenal gland problems. UCM 36244: 26 July 2013. Available at: www.fda.gov/downloads/Drugs/DrugSafety/UCM362444.pdf. Accessed May 12, 2015.

      • Raschi E.
      • Poluzzi E.
      • Koci A.
      • Caraceni P.
      • Ponti F.D.
      Assessing liver injury associated with antimycotics: Concise literature review and clues from data mining of the FAERS database.
      Few studies have examined the relative and absolute risks of acute liver injury associated with use of oral azole antifungals in clinical practice, and none have evaluated laboratory tests of liver inflammation or function. Moreover, it remains unclear whether chronic liver disease increases the risk of azole-associated acute liver injury. These data are important to differentiate azoles with little likelihood for acute liver injury from those with increased potential for this outcome.
      We evaluated the absolute and comparative risks of acute liver injury associated with oral azole antifungal drugs by examining incident elevations in liver aminotransferase levels and development of hepatic dysfunction among new initiators of these drugs in the outpatient setting. We also evaluated whether azole users with pre-existing chronic liver disease had a higher risk of acute liver injury than users without underlying liver disease.

      Methods

      Study Design and Data Source

      We conducted a retrospective cohort study using data from Kaiser Permanente Northern California (KPNC), an integrated healthcare organization that provides inpatient and outpatient services to Northern California residents.
      • Friedman G.
      • Habel L.
      • Boles M.
      • McFarland B.
      Kaiser Permanente Medical Care Program: Division of Research, Northern California, and Center for Health Research, Northwest Division.
      Data collected by KPNC included demographics; outpatient and hospital International Classification of Diseases, Ninth Revision (ICD-9) diagnoses; procedures; inpatient and outpatient laboratory results; emergency and referral services at non-Kaiser Permanente facilities; dispensed medications, including dosage, administration, and days' supply; and death date. Prescription drug benefits are utilized by >90% of members, and prior analyses have established the accuracy of pharmacy data.
      • Schatz M.
      • Zeiger R.S.
      • Vollmer W.M.
      • et al.
      Validation of a beta-agonist long-term asthma control scale derived from computerized pharmacy data.
      The study was approved by the KPNC and University of Pennsylvania institutional review boards.

      Study Patients

      Patients were eligible if they (1) newly initiated an oral azole (ie, fluconazole, ketoconazole, itraconazole, voriconazole, or posaconazole) in an outpatient setting between January 1, 2004 and December 31, 2010, (2) were ≥18 years old, and (3) were continuously enrolled in KPNC for ≥1 year before azole initiation. The index date represented the date the oral azole was first dispensed in an outpatient setting on or after January 1, 2004.
      Patients were excluded if, within 1 year before the index date, they were dispensed an azole in an outpatient setting, received warfarin (preventing identification of coagulopathy due to severe acute liver injury), or had evidence of severe acute liver injury (defined below). Patients prescribed more than 1 azole on the index date were also excluded.
      The baseline period was the 1 year before the index date. Follow-up continued until (1) study endpoint, (2) death, (3) disenrollment from KPNC, (4) switch to a different azole, (5) cessation of azole use (ie, no further fills within 30 days after the last prescription's days' supply), (6) dispensation of warfarin, or (7) last contact before December 30, 2010, whichever occurred first. For patients who discontinued azole use, we included 30 additional days of exposure time after the last days' supply to identify hepatotoxic events potentially related to azole use.
      • Chalasani N.P.
      • Hayashi P.H.
      • Bonkovsky H.L.
      • Navarro V.J.
      • Lee W.M.
      • Fontana R.J.
      ACG Clinical Guideline: the diagnosis and management of idiosyncratic drug-induced liver injury.
      For patients prescribed multiple courses of azoles, only the first course was evaluated.

      Main Study Outcomes

      To determine the full spectrum of acute liver injury events associated with oral azole use, we examined 3 outcomes. First, we determined development of severe liver aminotransferase elevations, defined as an inpatient or outpatient alanine aminotransferase (ALT) or aspartate aminotransferase (AST) >200 U/L (approximately 5 times the upper limit of normal of the assays used, a threshold that represents clinically important hepatic injury,
      • Aithal G.P.
      • Watkins P.B.
      • Andrade R.J.
      • et al.
      Case definition and phenotype standardization in drug-induced liver injury.
      and approximately 10 times what has been considered normal liver aminotransferase levels for males [30 U/L] and females [19 U/L]
      • Prati D.
      • Taioli E.
      • Zanella A.
      • et al.
      Updated definitions of healthy ranges for serum alanine aminotransferase levels.
      ). Next we evaluated severe acute liver injury, defined by an inpatient or outpatient international normalized ratio (INR) ≥1.5 and total bilirubin >2 times the upper limit of normal within 30 days of each other.
      • Lo Re 3rd, V.
      • Haynes K.
      • Goldberg D.
      • et al.
      Validity of diagnostic codes to identify cases of severe acute liver injury in the US Food and Drug Administration's Mini-Sentinel Distributed Database.
      This definition indicates severe hepatic dysfunction and has been used by the FDA's Mini-Sentinel initiative to assess serious drug-induced hepatotoxicity in the postmarketing setting.
      • Lo Re 3rd, V.
      • Haynes K.
      • Goldberg D.
      • et al.
      Validity of diagnostic codes to identify cases of severe acute liver injury in the US Food and Drug Administration's Mini-Sentinel Distributed Database.
      Given the inherent clinical challenges in confirming a drug-induced etiology for acute liver injury,
      • Rockey D.C.
      • Seeff L.B.
      • Rochon J.
      • et al.
      Causality assessment in drug-induced liver injury using a structured expert opinion process: comparison to the Roussel-Uclaf causality assessment method.
      we evaluated these 2 acute liver injury endpoints, regardless of etiology, to enable determination of the relative incidences of these outcomes among users of each azole. Finally, we determined acute liver failure events among azole initiators without chronic liver disease, because pre-existing liver disease precludes a diagnosis of acute liver failure.
      • Polson J.
      • Lee W.M.
      AASLD position paper: the management of acute liver failure.
      Acute liver failure was confirmed if a patient was hospitalized and had (1) no chronic liver disease, (2) coagulopathy (INR ≥1.5) without anticoagulation therapy, and (3) either hepatic encephalopathy or liver transplantation for acute liver failure.
      • Polson J.
      • Lee W.M.
      AASLD position paper: the management of acute liver failure.

      Data Collection

      Baseline data included age, sex, race, ethnicity, obesity (body mass index >30 kg/m2), alcohol dependence/abuse, cancer (excluding nonmelanoma skin cancers), chronic liver disease, diabetes mellitus, heart failure, human immunodeficiency virus (HIV) infection, and indication for azole prescription. Alcohol dependence/abuse,
      • Justice A.C.
      • Lasky E.
      • McGinnis K.A.
      • et al.
      Medical disease and alcohol use among veterans with human immunodeficiency infection: a comparison of disease measurement strategies.
      heart failure,
      • Saczynski J.S.
      • Andrade S.E.
      • Harrold L.R.
      • et al.
      A systematic review of validated methods for identifying heart failure using administrative data.
      HIV infection,
      • Fultz S.L.
      • Skanderson M.
      • Mole L.A.
      • et al.
      Development and verification of a “virtual” cohort using the National VA Health Information System.
      and chronic liver disease (Appendix 1A, available online) were defined by ICD-9 diagnoses. Chronic liver disease status was defined dichotomously (present vs absent). Diabetes and cancer were determined by KPNC registries.
      Outpatient and inpatient ALT, AST, INR, and total bilirubin results measured during follow-up were collected to assess liver aminotransferases >200 U/L and severe acute liver injury.
      Acute liver failure events were confirmed using a method that we have previously described.
      • Lo Re 3rd, V.
      • Carbonari D.M.
      • Forde K.A.
      • et al.
      Validity of diagnostic codes and laboratory tests of liver dysfunction to identify acute liver failure events.
      • Goldberg D.S.
      • Forde K.A.
      • Carbonari D.M.
      • et al.
      Population-representative incidence of drug-induced acute liver failure based on an analysis of an integrated health care system.
      Patients without chronic liver disease were screened for a potential acute liver failure event if, during follow-up, they had (1) a hospital ICD-9 diagnosis suggestive of acute liver failure (Appendix 1B, available online), and (2) an inpatient INR ≥1.5 and peak total bilirubin ≥5.0 mg/dL. Hospital records of potential acute liver failure patients (reviewed for all data in Appendix 2, available online) were abstracted onto structured forms that were then independently reviewed by 2 hepatologists. Disagreements were arbitrated by a third hepatologist. Determination of whether an azole was the cause of acute liver failure was based on consensus opinion by the hepatologists.

      Statistical Analysis

      For each cohort of azole initiators, we determined absolute risks and incidence rates (events per 1000 person-years) of endpoints with 95% confidence intervals (CIs). Given the potential for chronic liver disease to alter the magnitude of the association between drugs and acute liver injury,
      • Lewis J.H.
      The rational use of potentially hepatotoxic medications in patients with underlying liver disease.
      • Gupta N.K.
      • Lewis J.H.
      Review article: the use of potentially hepatotoxic drugs in patients with liver disease.
      we stratified outcomes by pre-existing chronic liver disease status. We determined whether incidence rates of outcomes were different by chronic liver disease status through a test of interaction using a Poisson regression model.
      Because azoles may be used to treat local or systemic fungal infections,
      • Como J.A.
      • Dismukes W.E.
      Oral azole drugs as systemic antifungal therapy.
      in a secondary analysis we grouped azoles by indication. We evaluated events among fluconazole, ketoconazole, and itraconazole users, because these are treatments for local fungal infections. We separately compared outcomes in fluconazole, voriconazole, and posaconazole users, because these are treatments for systemic fungal infections. Fluconazole appears in both groups because it can be used for local or systemic antifungal treatment.
      • Como J.A.
      • Dismukes W.E.
      Oral azole drugs as systemic antifungal therapy.
      Cox regression was used to estimate adjusted hazard ratios (HRs) of endpoints associated with use of each azole within an indication group.
      • Collett D.
      Modelling Survival Data in Medical Research.
      Fluconazole, the most frequently used azole, was the reference. Variables in Table 1 were evaluated as confounders. Proportionality of hazards was assessed by Schoenfeld residuals.
      • Hosmer D.W.
      • Lemeshow S.
      Applied Survival Analysis: Regression Modeling of Time-to-Event Data.
      For the analysis evaluating severe liver aminotransferase elevations, azole initiators who had an ALT or AST >200 U/L during the baseline period were excluded. For the analysis examining severe acute liver injury, there were many potential confounders relative to the number of events. Thus, for this analysis, we developed propensity scores, using logistic regression, to control for these variables, with fluconazole as the dependent variable.
      • Braitman L.E.
      • Rosenbaum P.R.
      Rare outcomes, common treatments: analytic strategies using propensity scores.
      • D'Agostino Jr., R.B.
      Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group.
      All variables in Table 1 were included in propensity score models. Cox regression was used to determine HRs of severe acute liver injury, adjusting for propensity score as a continuous variable. Acute liver failure was too rare to evaluate with Cox regression.
      Table 1Baseline Characteristics of New Initiators of an Oral Azole Antifungal Medication Dispensed in an Outpatient Setting Within the Kaiser Permanente Northern California Healthcare System (2004-2010), Overall and by Azole Drug
      CharacteristicOverall (n = 195,334)Fluconazole (n = 178,879)Ketoconazole (n = 14,296)Itraconazole (n = 1653)Voriconazole (n = 478)Posaconazole (n = 28)
      Age (y), median, IQR41 (30-54)41 (30-54)40 (28-53)50 (38-60)59 (50-67)55 (49-60)
      Female sex162,249 (83.1)155,730 (87.1)5539 (38.8)761 (46.0)211 (44.1)8 (28.6)
      Race
       White100,354 (51.4)91,916 (51.4)7177 (50.2)944 (57.1)298 (62.3)19 (67.9)
       Asian18,440 (9.4)16,852 (9.4)1375 (9.6)145 (8.7)64 (13.4)4 (14.3)
       Black or African American21,315 (10.9)19,701 (11.0)1472 (10.3)106 (6.4)34 (7.1)2 (7.1)
       Native Hawaiian/Other Pacific Islander705 (0.4)618 (0.4)81 (0.6)6 (0.4)0 (0.0)0 (0.0)
       American Indian/Alaska Native1213 (0.6)1114 (0.6)88 (0.6)6 (0.4)5 (1.1)0 (0.0)
       Unknown53,307 (27.3)48,678 (27.2)4103 (28.7)446 (27.0)77 (16.1)3 (10.7)
      Hispanic38,253 (25.8)35,757 (26.0)2142 (22.6)292 (22.9)59 (13.9)3 (12.5)
      Body mass index ≥30 kg/m265,886 (39.0)60,841 (39.1)4409 (38.8)511 (37.0)115 (26.9)10 (50.0)
      Diabetes mellitus21,747 (11.1)20,301 (11.4)1154 (8.1)159 (9.6)123 (25.7)10 (35.7)
      Cancer15,510 (7.9)14,087 (7.9)1061 (7.4)127 (7.7)219 (45.8)16 (57.1)
      Chronic liver disease7631 (3.9)7073 (4.0)397 (2.8)55 (3.3)97 (20.3)9 (32.1)
      Heart failure3679 (1.9)3416 (1.9)170 (1.2)38 (2.3)52 (10.9)3 (10.7)
      History of alcohol dependence/abuse36,741 (18.8)34,072 (19.1)2322 (16.2)242 (14.6)97 (20.3)8 (28.6)
      HIV infection1094 (0.6)1003 (0.6)57 (0.4)28 (1.7)6 (1.3)0 (0.0)
      Indication for prescription
       Systemic fungal infection16,536 (8.5)12,608 (7.1)3326 (23.3)382 (23.1)204 (42.7)16 (57.1)
       Vaginal candidiasis11,845 (6.1)11,727 (6.6)96 (0.7)14 (0.9)8 (1.7)0 (0.0)
       Onychomycosis6174 (3.12)5225 (2.9)481 (3.4)452 (27.3)15 (3.1)1 (3.6)
       Prostate cancer1160 (0.6)582 (0.3)560 (3.9)13 (0.8)5 (1.1)0 (0.0)
      Values are number (percentage) unless otherwise noted.
      HIV = human immunodeficiency virus; IQR = interquartile range.
      We also used Cox regression to determine whether chronic liver disease was an independent risk factor for liver aminotransferases >200 U/L and severe acute liver injury among all azole initiators. Variables in Table 1 were evaluated as potential confounders.
      As a sensitivity analysis, we repeated analyses with follow-up extended to 182 days after the last azole fill's days' supply, because drug-induced liver injury has been reported up to 6 months after drug discontinuation.
      • Chalasani N.P.
      • Hayashi P.H.
      • Bonkovsky H.L.
      • Navarro V.J.
      • Lee W.M.
      • Fontana R.J.
      ACG Clinical Guideline: the diagnosis and management of idiosyncratic drug-induced liver injury.
      Data were analyzed using SAS 9.4 (SAS Institute, Cary, NC).

      Results

      Patient Characteristics

      Among 5,484,224 KPNC members, 195,334 initiated an oral azole and met eligibility criteria (178,879 fluconazole; 14,296 ketoconazole; 1653 itraconazole; 478 voriconazole; 28 posaconazole [Figure]). The median days' supply prescribed was 2 days for fluconazole, 10 days for ketoconazole, 21 days for itraconazole, 30 days for voriconazole, and 29 days for posaconazole. Appendices 3A and 3B (available online) report reasons for censoring.
      Figure thumbnail gr1
      FigureSelection of patients for the study. *Azole antifungal agents of interest include ketoconazole, itraconazole, fluconazole, voriconazole, and posaconazole. †Index date defined as earliest qualifying date an oral antifungal drug was dispensed in an outpatient setting on or after January 1, 2004. ‡Severe acute liver injury defined by inpatient or outpatient international normalized ratio ≥1.5 (in the absence of anticoagulation therapy) and serum total bilirubin >2 times upper limit of normal, with both abnormalities recorded within 30 days of each other. KPNC = Kaiser Permanente Northern California.
      Voriconazole and posaconazole initiators were older, more commonly white or Asian, less commonly obese, and more frequently diagnosed with cancer, chronic liver disease, diabetes, heart failure, and a history of alcohol dependence/abuse compared with other azole initiators (Table 1). Fluconazole users were more commonly female. Onychomycosis was the most common indication for itraconazole initiators, whereas systemic fungal infections were the most frequent indication recorded for fluconazole, voriconazole, and posaconazole users.

      Liver Aminotransferases >200 U/L

      After excluding 29 patients with baseline transaminases >200 U/L, 336 (0.2%) developed an ALT or AST >200 U/L after azole initiation (Table 2). The median time to this outcome from initiation was 23 days (interquartile range, 10-38 days). We observed 265 events among fluconazole initiators (1 of 675 users; 13 events per 1000 person-years), 40 among ketoconazole initiators (1 of 357 users; 19.3 events per 1000 person-years), 8 among itraconazole initiators (1 of 207 users; 24.5 events per 1000 person-years), 21 among voriconazole initiators (1 of 23 users; 181.9 events per 1000 person-years), and 2 among posaconazole initiators (1 of 14 users; 191.1 events per 1000 person-years). Among the 336 patients who developed aminotransferases >200 U/L, 5 (1.5%; 3 fluconazole, 1 posaconazole, 1 voriconazole) were hospitalized for acute liver injury within 6 months of the event, and 14 (4.2%; 9 fluconazole, 1 itraconazole, 2 ketoconazole, 2 voriconazole) died within 6 months.
      Table 2Cumulative Incidence, Incidence Rate, and Risk of Liver Aminotransferases >200 U/L Among Initiators of Oral Azole Antifungal Agents in Kaiser Permanente Northern California, 2004-2010
      DrugNo. ExposedNo. Person-YearsNo. EventsCumulative Incidence per 1000 Persons (95% CI)Incidence Rate, Events/1000 Person-Years (95% CI)Unadjusted Hazard Ratio (95% CI) of ALT/AST >200 U/LAdjusted Hazard Ratio (95% CI) of ALT/AST >200 U/L
      Adjusted for age, sex, diabetes mellitus, cancer, chronic liver disease, heart failure, history of alcohol dependence/abuse, HIV infection, and indication for azole prescription.
      Outcomes among fluconazole, ketoconazole, and itraconazole users
       Fluconazole178,85220,4422651.5 (1.3-1.7)13.0 (11.4-14.6)Ref.Ref.
      No chronic liver disease171,78919,5771821.1 (0.9-1.2)9.3 (8.0-10.7)Ref.Ref.
      Chronic liver disease70638658311.8 (9.4-14.6)95.9 (76.4-118.9)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.001.
      Ref.Ref.
       Ketoconazole14,2942075402.8 (2.0-3.8)19.3 (13.8-26.3)1.50 (1.07-2.09)0.86 (0.59-1.27)
      No chronic liver disease13,8972010372.7 (1.9-3.7)18.4 (13.0-25.4)1.97 (1.38-2.82)1.01 (0.67-1.53)
      Chronic liver disease3976537.6 (1.6-22.1)46.1 (9.5-134.8)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.001.
      0.56 (0.18-1.77)0.51 (0.15-1.69)
       Itraconazole165332784.8 (2.1-9.5)24.5 (10.6-48.2)1.93 (0.95-3.91)1.59 (0.78-3.24)
      No chronic liver disease159831785.0 (2.2-9.9)25.3 (10.9-49.8)2.72 (1.33-5.56)2.19 (1.07-4.51)
      Chronic liver disease551000.0 (0.0-67.1)0.0 (0.0-356.6)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Outcomes among fluconazole, voriconazole, and posaconazole users
       Fluconazole178,85220,4422651.5 (1.3-1.7)13.0 (11.4-14.6)Ref.Ref.
      No chronic liver disease171,78919,5771821.1 (0.9-1.2)9.3 (8.0-10.7)Ref.Ref.
      Chronic liver disease70638658311.8 (9.4-14.6)95.9 (76.4-118.9)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.001.
      Ref.Ref.
       Voriconazole4781152143.9 (27.2-67.2)181.9 (112.6-278.0)15.9 (10.0-25.2)3.8 (2.4-6.1)
      No chronic liver disease381921744.6 (26.0-71.4)185.6 (108.1-297.2)22.7 (13.5-38.1)5.4 (3.1-9.2)
      Chronic liver disease9724441.2 (11.2-105.6)167.4 (45.6-428.5)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      2.2 (0.8-6.2)1.6 (0.6-4.6)
       Posaconazole2810271.4 (8.7-258.0)191.1 (23.1-690.4)19.4 (4.7-79.0)3.8 (0.9-15.6)
      No chronic liver disease1992105.3 (12.7-380.2)227.7 (27.6-822.6)2.8 (7.9-135.9)6.9 (1.7-29.0)
      Chronic liver disease9200.0 (0.0-409.9)0.0 (0.0-2194)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.001.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Azoles are grouped by indications. Outcomes among fluconazole, ketoconazole, and itraconazole users are compared in top half of table, because these drugs are prescribed for local fungal infection treatment. Outcomes in fluconazole, voriconazole, and posaconazole users are compared in bottom half, because these drugs are prescribed for systemic fungal infection treatment.
      ALT = alanine aminotransferase; AST = aspartate aminotransferase; CI = confidence interval.
      Adjusted for age, sex, diabetes mellitus, cancer, chronic liver disease, heart failure, history of alcohol dependence/abuse, HIV infection, and indication for azole prescription.
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.001.
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      § Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      After adjustment for age, sex, diabetes, cancer, chronic liver disease, heart failure, history of alcohol dependence/abuse, HIV infection, and indication for azole prescription, risk of aminotransferases >200 U/L was not significantly higher with use of ketoconazole, itraconazole, or posaconazole than fluconazole (Table 2). Risks of this outcome were similarly increased with use of voriconazole and posaconazole compared with fluconazole, but only reached statistical significance with voriconazole. Among users without chronic liver disease, the risk of this outcome was higher among users of itraconazole, voriconazole, and posaconazole than fluconazole (Table 2). Extending follow-up to 182 days after the last azole prescription's days' supply yielded similar findings (Appendix 4, available online).

      Severe Acute Liver Injury

      Fifty (0.03%) severe acute liver injury events occurred during follow-up (Table 3), with a median time from azole initiation of 22 days (interquartile range, 9-32 days). We observed 41 severe acute liver injury events among fluconazole initiators (1 of 4363 users; 2.0 events per 1000 person-years), 6 among ketoconazole initiators (1/2,383 users; 2.9 events per 1000 person-years), 0 among itraconazole initiators (0 events per 1000 person-years), 2 among voriconazole initiators (1 of 239 users; 16.7 events per 1000 person-years), and 1 among posaconazole initiators (1 of 28 users; 93.4 events per 1000 person-years). Among the 50 patients who developed severe acute liver injury, 1 was subsequently hospitalized for acute liver injury (ketoconazole) within 6 months of the event, and 9 (18.0%; 7 fluconazole, 1 ketoconazole, 1 posaconazole) died within 6 months.
      Table 3Cumulative Incidence, Incidence Rate, and Risk of Severe Acute Liver Injury Among Initiators of Oral Azole Antifungal Agents in Kaiser Permanente Northern California, 2004-2010
      DrugNo. ExposedNo. Person-YearsNo. Events
      Severe acute liver injury defined by inpatient or outpatient international normalized ratio ≥1.5 and serum total bilirubin >2 times upper limit of normal, with both abnormalities recorded within 30 days of each other. The upper limit of normal was determined by the assay from which the result was measured.
      Cumulative Incidence per 1000 Persons (95% CI)Incidence Rate, Events/1000 Person-Years (95% CI)Unadjusted Hazard Ratio (95% CI) of Severe Acute Liver Injury
      Severe acute liver injury defined by inpatient or outpatient international normalized ratio ≥1.5 and serum total bilirubin >2 times upper limit of normal, with both abnormalities recorded within 30 days of each other. The upper limit of normal was determined by the assay from which the result was measured.
      Propensity Score-Adjusted Hazard Ratio (95% CI) of Severe Acute Liver Injury
      Severe acute liver injury defined by inpatient or outpatient international normalized ratio ≥1.5 and serum total bilirubin >2 times upper limit of normal, with both abnormalities recorded within 30 days of each other. The upper limit of normal was determined by the assay from which the result was measured.
      Outcomes among fluconazole, ketoconazole, and itraconazole users
       Fluconazole178,87920,472410.2 (0.2-0.3)2.0 (1.4-2.7)Ref.Ref.
      No chronic liver disease171,80619,596230.1 (0.1-0.2)1.2 (0.7-1.8)Ref.Ref.
      Chronic liver disease7073877182.5 (1.5-4.0)20.5 (12.2-32.5)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.001.
      Ref.Ref.
       Ketoconazole14,296208260.4 (0.2-0.9)2.9 (1.1-6.3)1.57 (0.67-3.72)0.95 (0.37-2.42)
      No chronic liver disease13,899201760.4 (0.2-0.9)3.0 (1.1-6.5)2.81 (1.14-6.93)1.44 (0.53-3.94)
      Chronic liver disease3976600.0 (0.0-9.3)0.0 (0.0-56.3)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.001.
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
       Itraconazole165332800.0 (0.0-2.2)0.0 (0.0.-11.2)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      No chronic liver disease159831800.0 (0.0-2.3)0.0 (0.0-11.6)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      Chronic liver disease551000.0 (0.0-67.1)0.0 (0.0-356.6)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      Outcomes among fluconazole, voriconazole, and posaconazole users
       Fluconazole178,87920,472410.2 (0.2-0.3)2.0 (1.4-2.7)Ref.Ref.
      No chronic liver disease171,80619,596230.1 (0.1-0.2)1.2 (0.7-1.8)Ref.Ref.
      Chronic liver disease7073877182.5 (1.5-4.0)20.5 (12.2-32.5)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.001.
      Ref.Ref.
       Voriconazole47812024.2 (0.5-15.1)16.7 (2.0-60.2)9.0 (2.1-39.2)4.8 (1.1-21.2)
      No chronic liver disease3819500.0 (0.0-9.7)0.0 (0.0-38.8)
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Chronic liver disease9725220.6 (2.5-74.5)80.2 (9.7-289.8)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      4.5 (0.97-20.7)3.7 (0.8-17.0)
       Posaconazole2811135.7 (0.9-199.0)93.4 (2.4-520.6)
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      No chronic liver disease19900.0 (0.0-194.2)0.0 (0.0-410.7)
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Chronic liver disease921111.1 (2.8-619.1)581.6 (14.7-3241)
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.001.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Azoles are grouped by indications. Outcomes among fluconazole, ketoconazole, and itraconazole users are compared in top half, because these are prescribed for local fungal infection treatment. Outcomes in fluconazole, voriconazole, and posaconazole users are compared in bottom half, because these drugs are prescribed for systemic fungal infection treatment.
      CI = confidence interval.
      Severe acute liver injury defined by inpatient or outpatient international normalized ratio ≥1.5 and serum total bilirubin >2 times upper limit of normal, with both abnormalities recorded within 30 days of each other. The upper limit of normal was determined by the assay from which the result was measured.
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.001.
      Difference in incidence rates between chronic liver disease and no chronic liver disease, P <.05.
      § Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      After propensity score adjustment, risk of severe acute liver injury was not increased with use of ketoconazole, but was higher with voriconazole, compared with fluconazole. Results were similar when follow-up was extended to 182 days after the last azole prescription's days' supply (Appendix 5, available online). There were too few events among itraconazole and posaconazole users to permit evaluation within multivariable models.

      Acute Liver Failure

      Among 187,703 azole initiators without chronic liver disease, 13 (0.01%; 11 fluconazole, 2 ketoconazole) were hospitalized with potential acute liver failure within 182 days after the last days' supply. One patient developed acute liver failure (incidence rate, 0.5 [95% CI, 0.01-2.68] events per 1000 person-years), and ketoconazole was the implicated drug. The event occurred 40 days after initiation, and the patient required liver transplantation. No acute liver failure events were identified among other cohorts of azole users.

      Influence of Chronic Liver Disease

      Users of fluconazole and ketoconazole with chronic liver disease had higher rates of liver aminotransferase levels >200 U/L than those without underlying liver disease (P value interaction <.001; Table 2). Among all azole initiators, after adjustment for age, sex, and cancer diagnosis, chronic liver disease was associated with an increased risk of aminotransferases >200 U/L (HR 4.68; 95% CI, 3.68-5.94).
      Incidence rates of severe acute liver injury were also higher among initiators of fluconazole, voriconazole, and posaconazole who had chronic liver disease (P value interaction <.001; Table 3). Among all azole initiators, after adjustment for age, sex, race, obesity, and cancer diagnosis, chronic liver disease remained a strong risk factor for severe acute liver injury (HR 5.62; 95% CI, 2.56-12.35).

      Discussion

      In this study, the absolute risks and rates of both liver aminotransferase levels >200 U/L and severe acute liver injury (manifested by hepatic dysfunction) were similar among fluconazole, ketoconazole, and itraconazole users. Furthermore, among the 187,703 azole users without chronic liver disease, acute liver failure, the most severe form of acute liver injury, was confirmed in only 1 patient, a user of ketoconazole, highlighting the rarity of this event. The findings from this population-based study using liver-associated laboratory tests to define acute liver injury contradict the analyses by the FDA and EMA that suggested that ketoconazole use was associated with a higher risk of acute liver injury than other azole antifungals. However, these agencies' decisions were based primarily on spontaneous adverse event reports

      European Medicines Agency. European Medicines Agency recommends suspension of marketing authorisations for oral ketoconazole. Benefit of oral ketoconazole does not outweigh risk of liver injury in fungal infections. EMA/458028/2013: 26 July 2013. Available at: www.ema.europa.eu/docs/en_GB/document_library/Press_release/2013/07/WC500146613.pdf. Accessed May 12, 2015.

      US Food and Drug Administration. FDA Drug Safety Communication: FDA limits usage of Nizoral (ketoconazole) oral tablets due to potentially fatal liver injury and risk of drug interactions and adrenal gland problems. UCM 36244: 26 July 2013. Available at: www.fda.gov/downloads/Drugs/DrugSafety/UCM362444.pdf. Accessed May 12, 2015.

      and prior analyses of acute liver injury diagnosis codes.
      • Garcia Rodriguez L.A.
      • Duque A.
      • Castellsague J.
      • Perez-Gutthann S.
      • Stricker B.H.
      A cohort study on the risk of acute liver injury among users of ketoconazole and other antifungal drugs.
      Absolute risks and rates of liver aminotransferases >200 U/L and severe acute liver injury were comparable for voriconazole and posaconazole, but point estimates of these events were higher than with fluconazole. In multivariable analyses, voriconazole use was associated with increased risks of both outcomes compared with fluconazole use. However, given the small numbers of users of these drugs and extremely few events, these results should be interpreted with caution. Further, because fluconazole may also be used for local antifungal therapy (eg, vaginal candidiasis),
      • Como J.A.
      • Dismukes W.E.
      Oral azole drugs as systemic antifungal therapy.
      persons prescribed voriconazole or posaconazole may not be entirely comparable with those dispensed fluconazole, even after statistical adjustment for potential confounders.
      We found that users of fluconazole and ketoconazole with a history of chronic liver disease had higher absolute risks and rates of subsequent liver aminotransferases >200 U/L and that initiators of fluconazole, voriconazole, and posaconazole with chronic liver disease had higher rates of severe acute liver injury than users of these drugs without underlying chronic liver disease. Among all azole users, chronic liver disease was a strong risk factor for development of acute liver injury. We could not determine whether the acute liver injury events observed in those with chronic liver disease were caused by the azole medication or were due to the natural history of the underlying liver disease. However, patients with pre-existing liver disease typically have underlying hepatic inflammation and fibrosis, which can alter drug pharmacokinetics and hepatic metabolism and may reduce ability to withstand hepatic insults, potentially placing these patients at increased risk of drug-induced liver injury.
      • Gupta N.K.
      • Lewis J.H.
      Review article: the use of potentially hepatotoxic drugs in patients with liver disease.
      Our results should be interpreted within the context of several limitations. First, analyses of posaconazole and voriconazole were based on small sample sizes with few liver injury events. Second, misclassification of drug exposure is possible. Azole adherence cannot be confirmed, but we expect misclassification to be minimal because this was determined from dispensing records. Further, although our use of an as-treated approach captured acute liver injury events among patients on treatment, censoring patients when they stopped therapy might introduce bias if treatment cessation depended on risk of liver injury and varied by azole.
      • Hernan M.A.
      • Hernandez-Diaz S.
      • Robins J.M.
      A structural approach to selection bias.
      We attempted to account for this by extending follow-up for 182 days, to capture events that might be diagnosed after treatment discontinuation; this did not change the primary findings. Finally, there is potential for confounding by indication,
      • Walker A.M.
      Quantitative studies of the risk of serious hepatic injury in persons using nonsteroidal antiinflammatory drugs.
      because patients were assigned to treatment by clinician choice. Certain azoles, particularly voriconazole and posaconazole, may be prescribed to patients with more severe systemic fungal infections or comorbidities (eg, malignancy) that may increase their hepatotoxic potential. Our grouping of azoles by indication and use of statistical adjustment (controlling for indication) and propensity scores attempted to account for this,
      • Petri H.
      • Urquhart J.
      Channeling bias in the interpretation of drug effects.
      though a sizable proportion of azole users did not have an indication recorded.
      In conclusion, risks of acute liver injury were similarly low among users of fluconazole, ketoconazole, and itraconazole. In the subgroup without chronic liver disease, rates of liver aminotransferases >200 U/L were increased with itraconazole, voriconazole, and posaconazole. The risk of acute liver injury was higher with voriconazole than fluconazole, but results were based on few users and events. Pre-existing chronic liver disease was a strong risk factor for development of acute liver injury among azole users, and if confirmed, should possibly lead to recommendations for screening liver function testing in these patients before use.

      Supplementary Data

      Appendix 1AList of International Classification of Diseases, Ninth Revision (ICD-9) Diagnosis Codes to Identify Exclusionary Pre-existing Liver Diseases and Potential Acute Liver Failure Events
      ICD-9 CodeDescription
      Alpha-1-antitrypsin deficiency
       273.4Alpha-1-antitrypsin deficiency
      Autoimmune hepatitis
       571.42Autoimmune hepatitis
      Non-specific cirrhosis
       571.5Cirrhosis of liver without mention of alcohol
      Primary biliary cirrhosis
       571.6Biliary cirrhosis
      Hepatic decompensation
       572.3Portal hypertension
       789.5Ascites
       789.59Other ascites
       567.0Peritonitis infectious diseases classified elsewhere
       567.2Other suppurative peritonitis
       567.23Spontaneous bacterial peritonitis
       567.8Other specified peritonitis
       567.9Unspecified peritonitis
       456.0Esophageal varices with bleeding
       456.1Esophageal varices without mention of bleeding
       456.20Esophageal varices with bleed diseases classified elsewhere
       456.21Esophageal varices without mention of bleeding diseases classified elsewhere
      Alcoholic liver disease
       571Chronic liver disease and cirrhosis
       571.0Alcoholic fatty liver
       571.1Acute alcoholic hepatitis
       571.2Alcoholic cirrhosis of liver
       571.3Unspecified alcoholic liver damage
      Hepatitis B
       070.20Viral hepatitis B with hepatic coma acute/unspecified without hepatitis delta
       070.21Viral hepatitis B with hepatic coma acute/unspecified with hepatitis delta
       070.22Viral hepatitis B with hepatic coma chronic without hepatitis delta
       070.23Viral hepatitis B with hepatic coma chronic with hepatitis delta
       070.30Viral hepatitis B without hepatic coma acute/unspecified without hepatitis delta
       070.3Hepatitis B without coma
       070.31Viral hepatitis B without hepatic coma acute/unspecified with hepatitis delta
       070.32Viral hepatitis B without hepatic coma chronic without hepatitis delta
       070.33Viral hepatitis B without mention hepatic coma chronic with hepatitis delta
       V02.61Hepatitis B carrier
      Hepatitis C
       070.41Acute hepatitis C with hepatic coma
       070.44Chronic hepatitis C with hepatic coma
       070.51Acute hepatitis C without mention hepatic coma
       070.54Chronic hepatitis C without mention hepatic coma
       070.70Unspecified viral hepatitis C without hepatic coma
       070.71Unspecified viral hepatitis C with hepatic coma
       V02.62Hepatitis C carrier
      Hepatitis D
       070.42Hepatitis delta without mention active hepatitis B disease with hepatic coma
       070.52Hepatitis delta without mention active hepatitis B disease/hepatic coma
      Hemochromatosis
       275.0Disorders of iron metabolism
      Nonspecific or other specified hepatitis
       070Viral hepatitis
       070.49Other specified viral hepatitis with hepatic coma
       070.5Viral hepatic not elsewhere classified without coma
       070.59Other specified viral hepatitis without mention hepatic coma
       070.6Unspecified viral hepatitis with hepatic coma
       070.9Unspecified viral hepatitis without mention hepatic coma
       571.40Unspecified chronic hepatitis
       571.41Chronic persistent hepatitis
       573.1Hepatitis in viral diseases classified elsewhere
       573.2Hepatitis other infectious diseases classified elsewhere
       V02.6Viral hepatitis carrier
       V02.69Other viral hepatitis carrier
      Cancer in the liver and biliary tree
       155.0Malignant neoplasm of liver primary
       155Malignant neoplasm of liver and intrahepatic bile ducts
       155.1Malignant neoplasm of intrahepatic bile ducts
       155.2Malignant neoplasm liver not specified as primary/secondary
       197.7Secondary malignant neoplasm of liver
       230.8Carcinoma in situ of liver and biliary system
      Wilson's disease
       275.1Disorders of copper metabolism
      Nonalcoholic fatty liver disease
       571.8Other chronic nonalcoholic liver disease
      Appendix 1BList of International Classification of Diseases, Ninth Revision (ICD-9) Diagnoses to Identify Potential Acute Liver Failure Events.
      ICD-9 CodeDescription
      570Acute/subacute hepatic necrosis
      572.2Hepatic coma
      572.4Hepatorenal syndrome
      572.8Liver disease sequelae
      573.3Toxic (noninfectious) hepatitis
      573.8Other liver disorder, chemical/drug-induced
      V42.7Liver transplant
      Appendix 2Data Collected From Hospital Records to Permit Determination of Acute Liver Failure
      Laboratory ResultsPhysician Progress Notes, Admission History/Examination, and Hospital Discharge Summaries
      Alanine aminotransferaseDiagnoses of:
      Albumin Acute (or fulminant) liver (or hepatic) failure; “shock liver”
      Alkaline phosphatase Hepatic encephalopathy or encephalopathy
      Alpha-fetoproteinCause(s) of acute liver failure
      Drug-, herbal-, or dietary supplement-induced liver injury; ischemic injury; acute viral hepatitis; or other cause.
      AmmoniaEvidence to support a diagnosis of encephalopathy
      Asterixis (liver flap), altered mentation or delirium, change/decline in mental status, decreases consciousness, incoherence, dysarthria, confusion/disorientation, amnesia, lethargy, seizure, or coma/unresponsive.
      Anti-kidney liver microsomal type 1 antibodyLiver transplantation
      Anti-mitochondrial antibodyPre-existing liver disease
      Anti-nuclear antibodyUse of mechanical ventilation
      Anti-smooth muscle antibody
      Aspartate aminotransferase
      Ceruloplasmin
      Creatinine
      FerritinBrain Imaging Reports
      Reports from computed tomography and magnetic resonance imaging studies of the brain.
      GlucoseDiagnoses of:
      Hepatitis A IgM Acute cerebral ischemia
      Hepatitis B DNA Acute stroke
      Hepatitis B surface antigen Cerebral edema
      Hepatitis B e antigen Intracranial/cerebral bleed
      Hepatitis C antibody Intracranial/cerebral mass
      Hepatitis C RNA Uncal herniation
      Hepatitis E antibody
      International normalized ratio
      Iron
      Platelet countMedications Dispensed
      SodiumMedications used to treat encephalopathy
      Drugs to treat hepatic encephalopathy included: lactulose, flumazenil, intravenous mannitol, methylprednisolone, pentobarbital, rifaximin, and thiopental.
      Total bilirubin
      Transferrin
      Drug-, herbal-, or dietary supplement-induced liver injury; ischemic injury; acute viral hepatitis; or other cause.
      Asterixis (liver flap), altered mentation or delirium, change/decline in mental status, decreases consciousness, incoherence, dysarthria, confusion/disorientation, amnesia, lethargy, seizure, or coma/unresponsive.
      Reports from computed tomography and magnetic resonance imaging studies of the brain.
      § Drugs to treat hepatic encephalopathy included: lactulose, flumazenil, intravenous mannitol, methylprednisolone, pentobarbital, rifaximin, and thiopental.
      Appendix 3AReasons for Censoring During Evaluation of Severe Acute Liver Injury, Overall and by Azole Antifungal Medication
      Reason for CensorOverall (n = 195,334)Fluconazole (n = 178,879)Ketoconazole (n = 14,296)Itraconazole (n = 1,653)Voriconazole (n = 478)Posaconazole (n = 28)
      Severe acute liver injury50 (0.03)41 (0.02)6 (0.04)0 (0)2 (0.4)1 (3.6)
      Death1371 (0.7)1208 (0.7)101 (0.7)11 (0.7)48 (10.0)3 (10.7)
      End of enrollment2445 (1.3)2178 (1.2)228 (1.6)32 (1.9)6 (1.3)1 (3.6)
      Switched azole503 (0.3)294 (0.2)146 (1.0)36 (2.2)26 (5.4)1 (3.6)
      Discontinued
      Discontinuation of azole defined as no further azole prescription fills within 30 days after the last azole prescription's days' supply.
      azole
      187,733 (96.1)172,218 (96.3)13,598 (95.1)1533 (92.7)362 (75.7)22 (78.6)
      Dispensed warfarin312 (0.2)272 (0.2)20 (0.1)8 (0.5)12 (2.5)0 (0)
      End of study data2920 (1.5)2668 (1.5)197 (1.4)33 (2.0)22 (4.6)0 (0)
      Severe acute liver injury defined by inpatient or outpatient international normalized ratio ≥1.5 (in the absence of anticoagulation therapy) and serum total bilirubin >2 times upper limit of normal, with both abnormalities recorded within 30 days of each other. Values are presented as number (percentage).
      Discontinuation of azole defined as no further azole prescription fills within 30 days after the last azole prescription's days' supply.
      Appendix 3BReasons for Censoring During Evaluation of Serum Alanine Aminotransferase or Aspartate Aminotransferase Level >200 U/L, Overall and by Azole Antifungal Medication
      Reason for CensorOverall (n = 195,305)Fluconazole (n = 178,852)Ketoconazole (n = 14,294)Itraconazole (n = 1,653)Voriconazole (n = 478)Posaconazole (n = 28)
      Severe acute liver injury336 (0.2)265 (0.2)40 (0.3)8 (0.5)21 (4.4)2 (7.1)
      Death1341 (0.7)1187 (0.7)99 (0.7)9 (0.5)43 (9.0)3 (10.7)
      End of enrollment2441 (1.3)2174 (1.2)228 (1.6)32 (1.9)6 (1.3)1 (3.6)
      Switched azole497 (0.3)288 (0.2)146 (1.0)36 (2.2)26 (5.4)1 (3.6)
      Discontinued
      Discontinuation of azole defined as further azole prescription fills within 30 days after the last azole prescription's days' supply. Values are presented as number (percentage).
      azole
      187,471 (96.0)172,006 (96.2)13,567 (94.9)1527 (92.4)350 (73.2)21 (75.0)
      Dispensed warfarin305 (0.2)268 (0.2)19 (0.1)8 (0.5)10 (2.1)0 (0)
      End of study data2914 (1.5)2664 (1.5)195 (1.4)33 (2.0)22 (4.6)0 (0)
      Discontinuation of azole defined as further azole prescription fills within 30 days after the last azole prescription's days' supply. Values are presented as number (percentage).
      Appendix 4Cumulative Incidence, Incidence Rates, and Risk of Liver Aminotransferases >200 U/L Among Initiators of Oral Azole Antifungal Agents in Kaiser Permanente Northern California (2004-2010), with Follow-Up Extended to 182 Days After the Last Azole Prescription Fill's Days' Supply
      DrugNo. ExposedNo. Person-YearsNo. Events
      Adjusted for age, sex, diabetes, cancer, chronic liver disease, heart failure, history of alcohol dependence/abuse, human immunodeficiency virus infection, and indication for azole prescription.
      Cumulative Incidence per 1000 Persons (95% CI)Incidence Rate, Events/1000 Person-Years (95% CI)Unadjusted Hazard Ratio (95% CI) of ALT/AST >200 U/LFully Adjusted Hazard Ratio (95% CI) of ALT/AST >200 U/L*
      Outcomes among fluconazole, ketoconazole, and itraconazole users
       Fluconazole178,85287,7885913.3 (3.0-3.6)6.7 (6.2-7.3)Ref.Ref.
      No CLD171,78984,5074392.6 (2.3-2.8)5.2 (4.7-5.7)Ref.Ref.
      CLD7063328115221.5 (18.2-25.2)46.3 (39.3-54.3)
      Difference in incidence rates between CLD and no CLD, P <.001.
      Ref.Ref.
       Ketoconazole14,2947363694.8 (3.8-6.1)9.4 (7.3-11.9)1.39 (1.08-1.79)0.90 (0.69-1.19)
      No CLD13,8977161634.5 (3.5-5.8)8.8 (6.8-11.3)1.68 (1.29-2.19)1.07 (0.79-1.43)
      CLD397203615.1 (5.5-32.9)29.6 (10.9-64.4)
      Difference in incidence rates between CLD and no CLD, P <.05.
      0.67 (0.29-1.51)0.53 (0.23-1.25)
       Itraconazole1653931127.3 (3.8-12.7)12.9 (6.7-22.5)1.93 (1.09-3.43)1.41 (0.79-2.51)
      No CLD1598900116.9 (3.4-12.3)12.2 (6.1-21.9)2.35 (1.29-4.28)1.72 (0.93-3.15)
      CLD5531118.2 (0.5-101.3)32.6 (0.8-181.5)0.75 (0.10-5.37)0.70 (0.10-5.06)
      Outcomes among fluconazole, voriconazole, and posaconazole users
       Fluconazole178,85287,7885913.3 (3.0-3.6)6.7 (6.2-7.3)Ref.Ref.
      No CLD171,78984,5074392.6 (2.3-2.8)5.2 (4.7-5.7)Ref.Ref.
      CLD7063328115221.5 (18.2-25.2)46.3 (39.3-54.3)
      Difference in incidence rates between CLD and no CLD, P <.001.
      Ref.Ref.
       Voriconazole4782372450.2 (32.2-74.7)101.1 (64.8-150.4)14.8 (9.8-22.3)3.5 (2.3-5.4)
      No CLD3811901949.9 (30.0-77.9)100.0 (60.2-156.1)18.6 (11.7-29.7)5.4 (3.3-8.7)
      CLD9747551.5 (16.7-120.3)105.7 (34.3-246.6)
      Difference in incidence rates between CLD and no CLD, P <.05.
      2.4 (0.99-6.0)1.5 (0.6-3.8)
       Posaconazole2818271.4 (8.7-258.0)112.0 (13.6-404.5)17.5 (4.4-70.7)3.0 (0.7-12.1)
      No CLD19132105.3 (12.7-380.2)151.4 (18.3-546.9)30.8 (7.6-124.7)8.8 (2.2-36.0)
      CLD9500.0 (0.0-409.9)0.0 (0.0-793.5)
      Difference in incidence rates between CLD and no CLD, P <.001.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Azoles are grouped by indications. Outcomes among fluconazole, ketoconazole, and itraconazole users are compared in top half, because these drugs are prescribed for local fungal infection treatment. Outcomes in fluconazole, voriconazole, and posaconazole users are compared in bottom half, because these drugs are prescribed for systemic fungal infection treatment.
      ALT = alanine aminotransferase; AST = aspartate aminotransferase; CI = confidence interval; CLD = chronic liver disease.
      Adjusted for age, sex, diabetes, cancer, chronic liver disease, heart failure, history of alcohol dependence/abuse, human immunodeficiency virus infection, and indication for azole prescription.
      Difference in incidence rates between CLD and no CLD, P <.001.
      Difference in incidence rates between CLD and no CLD, P <.05.
      § Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Appendix 5Cumulative Incidence, Incidence Rates, and Risk of Severe Acute Liver Injury Among Initiators of Oral Azole Antifungal Agents in Kaiser Permanente Northern California (2004-2010), with Follow-Up Extended to 182 Days After the Last Azole Prescription Fill's Days' Supply
      DrugNo. ExposedNo. Person-YearsNo. EventsCumulative Incidence per 1,000 Persons (95% CI)Incidence Rate, Events/1,000 Person-Years (95% CI)Unadjusted Hazard Ratio (95% CI) of Severe Acute Liver InjuryPropensity Score-Adjusted Hazard Ratio (95% CI) of Severe Acute Liver Injury
      Outcomes among fluconazole, ketoconazole, and itraconazole users
       Fluconazole178,87987,952870.5 (0.4-0.6)1.0 (0.8-1.2)Ref.Ref.
      No CLD171,80684,628460.3 (0.2-0.4)0.5 (0.4-0.7)Ref.Ref.
      CLD70733324415.8 (4.2-7.9)12.3 (8.9-16.7)
      Difference in incidence rates between CLD and no CLD, P <.001.
      Ref.Ref.
       Ketoconazole14,2967387100.7 (0.3-1.3)1.4 (0.6-2.5)1.33 (0.69-2.56)0.81 (0.40-1.66)
      No CLD13,899718280.6 (0.2-1.1)1.1 (0.5-2.2)2.04 (0.96-4.34)0.95 (0.42-2.18)
      CLD39720425.0 (0.6-18.2)9.8 (1.2-35.4)0.76 (0.18-3.17)0.63 (0.14-2.79)
       Itraconazole165393500.0 (0.0-2.2)0.0 (0.0-3.9)
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      No CLD159890400.0 (0.0-2.3)0.0 (0.0-4.1)
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      CLD553100.0 (0.0-67.1)0.0 (0.0-120.0)
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Outcomes among fluconazole, voriconazole, and posaconazole users
       Fluconazole178,87987,952870.5 (0.4-0.6)1.0 (0.8-1.2)Ref.Ref.
      No CLD171,80684,628460.3 (0.2-0.4)0.5 (0.4-0.7)Ref.Ref.
      CLD70733324415.8 (4.2-7.9)12.3 (8.9-16.7)
      Difference in incidence rates between CLD and no CLD, P <.001.
      Ref.Ref.
       Voriconazole47824648.4 (2.3-21.4)16.3 (4.4-41.6)13.7 (4.9-38.1)7.3 (2.6-20.6)
      No CLD38119725.2 (0.6-19.0)10.1 (1.2-36.6)17.4 (4.1-73.2)4.6 (0.97-21.6)
      CLD9749220.6 (2.5-74.5)40.9 (5.0-147.9)
      Difference in incidence rates between CLD and no CLD, P <.05.
      2.8 (0.65-12.0)2.2 (0.5-9.6)
       Posaconazole2818135.7 (0.9-199.0)55.2 (1.4-307.6)
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      No CLD191400.0 (0.0-194.2)0.0 (0.0-267.2)
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      CLD941111.1 (2.8-619.1)232.1 (5.9-1293)
      Difference in incidence rates between CLD and no CLD, P <.001.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.
      Azoles are grouped by indications. Outcomes among fluconazole, ketoconazole, and itraconazole users are compared in top half, because these drugs are prescribed for local fungal infection treatment. Outcomes in fluconazole, voriconazole, and posaconazole users are compared in bottom half, because these drugs are prescribed for systemic fungal infection treatment. Severe acute liver injury defined by inpatient or outpatient international normalized ratio ≥1.5 and serum total bilirubin >2 times upper limit of normal, with both abnormalities recorded within 30 days of each other.
      CI = confidence interval; CLD = chronic liver disease.
      Difference in incidence rates between CLD and no CLD, P <.001.
      Difference in incidence rates between CLD and no CLD, P <.05.
      Hazard ratios were not determined because event rates and/or sample sizes were too small for analysis.

      References

        • Como J.A.
        • Dismukes W.E.
        Oral azole drugs as systemic antifungal therapy.
        N Engl J Med. 1994; 330: 263-272
        • Kim W.
        • Ryan C.J.
        Androgen receptor directed therapies in castration-resistant metastatic prostate cancer.
        Curr Treat Options Oncol. 2012; 13: 189-200
        • Lo Re 3rd, V.
        • Haynes K.
        • Goldberg D.
        • et al.
        Validity of diagnostic codes to identify cases of severe acute liver injury in the US Food and Drug Administration's Mini-Sentinel Distributed Database.
        Pharmacoepidemiol Drug Saf. 2013; 22: 861-872
        • Chalasani N.
        • Fontana R.J.
        • Bonkovsky H.L.
        • et al.
        Causes, clinical features, and outcomes from a prospective study of drug-induced liver injury in the United States.
        Gastroenterology. 2008; 135 (1934.e1-4): 1924-1934
      1. European Medicines Agency. European Medicines Agency recommends suspension of marketing authorisations for oral ketoconazole. Benefit of oral ketoconazole does not outweigh risk of liver injury in fungal infections. EMA/458028/2013: 26 July 2013. Available at: www.ema.europa.eu/docs/en_GB/document_library/Press_release/2013/07/WC500146613.pdf. Accessed May 12, 2015.

      2. US Food and Drug Administration. FDA Drug Safety Communication: FDA limits usage of Nizoral (ketoconazole) oral tablets due to potentially fatal liver injury and risk of drug interactions and adrenal gland problems. UCM 36244: 26 July 2013. Available at: www.fda.gov/downloads/Drugs/DrugSafety/UCM362444.pdf. Accessed May 12, 2015.

        • Raschi E.
        • Poluzzi E.
        • Koci A.
        • Caraceni P.
        • Ponti F.D.
        Assessing liver injury associated with antimycotics: Concise literature review and clues from data mining of the FAERS database.
        World J Hepatol. 2014; 6: 601-612
        • Friedman G.
        • Habel L.
        • Boles M.
        • McFarland B.
        Kaiser Permanente Medical Care Program: Division of Research, Northern California, and Center for Health Research, Northwest Division.
        in: Strom B.L. Pharmacoepidemiology. 3rd ed. John Wiley & Sons, Ltd, West Sussex2000: 263-283
        • Schatz M.
        • Zeiger R.S.
        • Vollmer W.M.
        • et al.
        Validation of a beta-agonist long-term asthma control scale derived from computerized pharmacy data.
        J Allergy Clin Immunol. 2006; 117: 995-1000
        • Chalasani N.P.
        • Hayashi P.H.
        • Bonkovsky H.L.
        • Navarro V.J.
        • Lee W.M.
        • Fontana R.J.
        ACG Clinical Guideline: the diagnosis and management of idiosyncratic drug-induced liver injury.
        Am J Gastroenterol. 2014; 109 (quiz 67): 950-966
        • Aithal G.P.
        • Watkins P.B.
        • Andrade R.J.
        • et al.
        Case definition and phenotype standardization in drug-induced liver injury.
        Clin Pharmacol Ther. 2011; 89: 806-815
        • Prati D.
        • Taioli E.
        • Zanella A.
        • et al.
        Updated definitions of healthy ranges for serum alanine aminotransferase levels.
        Ann Intern Med. 2002; 137: 1-10
        • Rockey D.C.
        • Seeff L.B.
        • Rochon J.
        • et al.
        Causality assessment in drug-induced liver injury using a structured expert opinion process: comparison to the Roussel-Uclaf causality assessment method.
        Hepatology. 2010; 51: 2117-2126
        • Polson J.
        • Lee W.M.
        AASLD position paper: the management of acute liver failure.
        Hepatology. 2005; 41: 1179-1197
        • Justice A.C.
        • Lasky E.
        • McGinnis K.A.
        • et al.
        Medical disease and alcohol use among veterans with human immunodeficiency infection: a comparison of disease measurement strategies.
        Med Care. 2006; 44: S52-S60
        • Saczynski J.S.
        • Andrade S.E.
        • Harrold L.R.
        • et al.
        A systematic review of validated methods for identifying heart failure using administrative data.
        Pharmacoepidemiol Drug Saf. 2012; 21: 129-140
        • Fultz S.L.
        • Skanderson M.
        • Mole L.A.
        • et al.
        Development and verification of a “virtual” cohort using the National VA Health Information System.
        Med Care. 2006; 44: S25-S30
        • Lo Re 3rd, V.
        • Carbonari D.M.
        • Forde K.A.
        • et al.
        Validity of diagnostic codes and laboratory tests of liver dysfunction to identify acute liver failure events.
        Pharmacoepidemiol Drug Saf. 2015; 24: 676-683
        • Goldberg D.S.
        • Forde K.A.
        • Carbonari D.M.
        • et al.
        Population-representative incidence of drug-induced acute liver failure based on an analysis of an integrated health care system.
        Gastroenterology. 2015; 148: 1353-1361.e3
        • Lewis J.H.
        The rational use of potentially hepatotoxic medications in patients with underlying liver disease.
        Expert Opin Drug Saf. 2002; 1: 159-172
        • Gupta N.K.
        • Lewis J.H.
        Review article: the use of potentially hepatotoxic drugs in patients with liver disease.
        Aliment Pharmacol Ther. 2008; 28: 1021-1041
        • Collett D.
        Modelling Survival Data in Medical Research.
        2nd ed. Chapman and Hall/CRC Press, Boca Raton, FL2003
        • Hosmer D.W.
        • Lemeshow S.
        Applied Survival Analysis: Regression Modeling of Time-to-Event Data.
        John Wiley & Sons, New York1999
        • Braitman L.E.
        • Rosenbaum P.R.
        Rare outcomes, common treatments: analytic strategies using propensity scores.
        Ann Intern Med. 2002; 137: 693-695
        • D'Agostino Jr., R.B.
        Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group.
        Stat Med. 1998; 17: 2265-2281
        • Garcia Rodriguez L.A.
        • Duque A.
        • Castellsague J.
        • Perez-Gutthann S.
        • Stricker B.H.
        A cohort study on the risk of acute liver injury among users of ketoconazole and other antifungal drugs.
        Br J Clin Pharmacol. 1999; 48: 847-852
        • Hernan M.A.
        • Hernandez-Diaz S.
        • Robins J.M.
        A structural approach to selection bias.
        Epidemiology. 2004; 15: 615-625
        • Walker A.M.
        Quantitative studies of the risk of serious hepatic injury in persons using nonsteroidal antiinflammatory drugs.
        Arthritis Rheum. 1997; 40: 201-208
        • Petri H.
        • Urquhart J.
        Channeling bias in the interpretation of drug effects.
        Stat Med. 1991; 10: 577-581