Author + information
- aBoston University School of Medicine, Boston, Massachusetts
- bDivision of Gastroenterology and Inborn Error Products, Center for Drug Evaluation and Research, Office of New Drugs/ODE III, U.S. Food and Drug Administration, Silver Spring, Maryland
- ↵∗Address for correspondence:
Dr. Elaine M. Hylek, Boston University School of Medicine, 801 Massachusetts Avenue, 2nd Floor Suite, Boston, Massachusetts 02118
- atrial fibrillation
- direct oral anticoagulant
- liver disease
- non-vitamin K oral antagonist
Similar to patients with advanced renal impairment, individuals with advanced liver disease constitute a very high-risk group for both serious bleeding and thromboembolism. Patients with moderate-to-severe hepatic dysfunction, defined as Child-Pugh class B or C (1), were excluded from the randomized trials of direct oral anticoagulants for stroke prevention in atrial fibrillation. The lack of data on safety and efficacy in this complex patient population is problematic, as vitamin K antagonist use has been associated with harm (2). Current prescribing information is cautionary for patients with hepatic impairment. For rivaroxaban and edoxaban, it is advised to “avoid use in patients with moderate or severe hepatic impairment (Child-Pugh B and C)” (3,4). For apixaban, the package insert states that “dosing recommendations cannot be provided for patients with moderate hepatic impairment (Child-Pugh class B), and is not recommended in patients with severe hepatic impairment (Child-Pugh class C)” (5). For dabigatran, “patients with elevated transaminases >2× upper limit of normal (ULN) were excluded in clinical trials; therefore, its use in this population is not recommended” (6).
In this issue of the Journal, Lee et al. (7) report their findings on the efficacy and safety of direct oral anticoagulants among individuals with atrial fibrillation and “significant active liver disease.” The authors define “significant active liver disease” as the subpopulation of subjects identified in their National Health Insurance Service database with administrative codes for liver cirrhosis, viral hepatitis, or abnormal alanine transaminase (ALT) or aspartate transaminase (AST) more than 2× ULN. This definition of significant active liver disease for prescribing clinicians is problematic. Furthermore, interpretation of these results is further complicated by the fact that 52.5% of patients were prescribed a reduced dosage of anticoagulant therapy.
The U.S. Food and Drug Administration Guidance for Industry for assessment of hepatic impairment for drug development recommends that “Child-Pugh classification be used to categorize the degree of hepatic impairment in patients, just as serum creatinine or creatinine clearance is used to categorize varying degrees of renal impairment” (8). Components of this classification include both laboratory and clinical parameters: bilirubin, albumin, prothrombin, encephalopathy and ascites. The Food and Drug Administration Guidance further states that “a study design involving control subjects and patients with a Child-Pugh category of moderate impairment would generally be appropriate. In that case, the findings in the moderate category would be applied to patients with a mild Child-Pugh category, and dosing in the severe category would generally be contraindicated.”
Owing to the lack of both clinical markers of severity as well as laboratory data available in the database as indicated by the authors, they were unable to calculate the Child-Pugh score and instead substituted the hepatocellular enzymes ALT and AST. Transaminases, however, do not provide functional information concerning hepatic metabolism or synthetic impairment. Exclusion of patients with elevated liver enzymes >2× ULN in the RE-LY (Randomized Evaluation of Long-Term Anticoagulation Therapy) trial with dabigatran was likely due to the need for heightened vigilance for hepatotoxicity as was experienced with the previous thrombin inhibitor ximelagatran, which was subsequently removed from the market (9,10). At a minimum, in the current study, inclusion of common laboratory parameters demonstrating evidence for hepatic impairment—increased direct bilirubin, hypoalbuminemia, and prolonged prothrombin time—would have fortified the authors’ assertions concerning the safety of the oral anticoagulants. In addition, the authors provide no information on the stability of these measures over time or the temporal association of the baseline measurement with initiation of anticoagulant therapy. Given that AST is found in many tissues including the liver, heart, skeletal muscle, kidneys, brain, and red blood cells, isolated elevation without concomitant elevation in bilirubin raises additional questions about misclassification. Of note, only 7% of patients included in the “significant active liver disease” group had a creatinine clearance <50 ml/min and the overall mean was approximately 81 ml/min reinforcing the question of an overall healthier cohort. Regarding degree of selection bias, it would be informative to know the proportion of patients meeting the same enrollment criteria that was not prescribed anticoagulant therapy.
Although the authors importantly highlight the increasing prevalence of nonalcoholic fatty liver disease (NAFLD), it is likely that such patients were represented in the randomized trials given the baseline prevalence of obesity and diabetes in the enrolled population. As an example, 40% of participants in the ARISTOTLE (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation) trial had a body mass index of ≥30 kg/m2 (11). However, it is unlikely that patients who had progressed to nonalcoholic steatohepatitis would have been adequately studied given the significantly lower prevalence of nonalcoholic steatohepatitis in the source population. Alterations in drug-metabolizing enzyme activity have been demonstrated in animal models of NAFLD, and as modulation of CYP3A4 in NAFLD is particularly relevant given its role in the metabolism of both apixaban and rivaroxaban, the authors raise an important question of ongoing active research regarding hepatic impairment in patients with hepatic steatosis (12).
In this study, only 768 patients were identified with cirrhosis, but due to the lack of clinical and laboratory information, it is not known how many had decompensated cirrhosis. Individuals with compensated cirrhosis do not have portal hypertension and do not have coagulopathy. It would have been most worthwhile to learn whether such patients (Child-Pugh class A patients) could safely take the direct oral anticoagulants. Decompensated cirrhotic patients are at significantly higher risk of developing coagulopathy and have portal hypertension. Anticoagulant drugs are generally contraindicated in patients with severe liver impairment, Child-Pugh class B and C, due to the presence of portal hypertension, portal gastropathy, varices, impaired renal function, thrombocytopenia, coagulopathy, baseline high risk of bleeding, and impaired drug metabolism.
Based on pharmacokinetic properties (less reliant on hepatic clearance, shorter half-life), one would anticipate that direct oral anticoagulants might be safer compared with warfarin for patients with hepatic impairment. The current study cannot address this question. The current study really provides not much new information than what most prescribers are likely to have known, given that the definition of severe liver disease was exceedingly broad. The authors did not provide any new evidence that these agents could be safely administered to Child-Pugh class A patients either. Research is needed with rigorously defined subgroups of patients with hepatic impairment across the spectrum of severity to best inform drug choice and dose.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
The opinions expressed are those of the authors and not of the U.S. Food and Drug Administration, the U.S. Department of Health and Human Services, or the U.S. Government. Dr. Hylek has received research support from Janssen; served as a consultant for Boehringer Ingelheim, Bristol-Myers Squibb/Pfizer, Janssen, Medtronic, and Roche; and has received lecture honoraria from Boehringer Ingelheim and Bristol-Myers Squibb/Pfizer. Dr. Anania has reported that he has no relationships relevant to the contents of this paper to disclose.
- Efird L.M.,
- Mishkin D.S.,
- Berlowitz D.R.,
- et al.
- Bayer Pharma A.G.
- Sankyo D.
- Bristol-Myers Squibb-Pfizer
- Boehringer-Ingelheim Pharmaceuticals
- Lee S.-R.,
- Lee H.-J.,
- Choi E.-K.,
- et al.
- U.S. Department of Health and Human Services Food and Drug Administration Center for Drug Evaluation and Research