Abstract
We report 2 patients with compensated cirrhosis and moderate renal impairment who experienced severe bleeding complications from paracentesis during concurrent therapy with apixaban. While paracentesis has traditionally been considered a low bleeding-risk procedure and safe to perform without interruption of therapeutic anticoagulation, the increased concentrations observed in patients with impaired liver function may place these patients at unexpectedly high bleeding risk. Further investigation into the safety of paracentesis in patients with cirrhosis on apixaban may be warranted, as well as additional understanding of the clinical safety of this drug in Child-Pugh B cirrhosis.
Case Presentations
We report 2 cases of unexpected major hemorrhage in patients undergoing routine large volume paracentesis in the context of combined moderate renal and moderate hepatic impairment, observed approximately 2 weeks apart.
Case 1
A 69-year-old man weighing 95.7 kg with Child-Pugh B cirrhosis, chronic kidney disease with baseline creatinine of 1.5 mg/dL and on apixaban 5 mg twice a day for atrial fibrillation presented for routine therapeutic paracentesis under ultrasound guidance for refractory ascites. Apixaban had been held for 24 hours prior to previous biweekly paracenteses; however, apixaban was not held prior to this index procedure and the patient’s last dose was approximately 2 hours prior to hospital arrival. Institutional guidance for low bleed risk procedures recommended a 24 hours apixaban hold. Paracentesis removed 3200 mL of straw-colored fluid. While initially no complications were identified, the patient developed symptomatic hypotension and syncope while waiting for his next clinic appointment. He presented to the emergency department 3.5 hours after paracentesis, and was profoundly hypotensive with blood pressures as low as 78/63 mmHg. On arrival to the ED, creatinine was 1.9 mg/dL, total bilirubin was 1.9 mg/dL and hemoglobin was 10.8 g/dL, down from 17.0 g/dL 5 days prior. No apixaban levels were drawn during this presentation. As an outpatient, intentional normalized ratio (INR) had recently been 1.3 to 1.5 but was 1.9 on ED presentation. A CT angiography of the abdomen and pelvis identified a focus of active bleeding along the peritoneal surface of the left lateral abdominal wall (Figure 1A). The patient was taken to intensive care, resuscitated, and received emergent interventional radiology arterial embolization. He received Factor Eight Inhibitor Bypass Activity (FEIBA). Unfortunately, blood pressures failed to respond and the patient died from complications of hemorrhagic shock 20 hours after paracentesis.
Both patients, Case 1 (Image (A)) and Case 2 (Image (B)), demonstrated small areas of contrast extravasation (circled areas) along the left anterior abdominal wall at the site of paracentesis on CT abdominal angiography, consistent with active bleeding at the puncture sites.
Case 2
A 75-year-old man weighing 81.6 kg with Child-Pugh B cirrhosis, chronic kidney disease and atrial fibrillation on apixaban was admitted to the hospital for weight gain and rising creatinine, which had increased from 2.0 to 3.0 mg/dL at the time of admission. Apixaban was initially continued at his home dose (5 mg by mouth twice a day) but was decreased to 2.5 mg by mouth twice a day starting hospital day 5. He was on aspirin prior to admission and was initiated on clopidogrel when a coronary stent was placed on hospital day 4. On hospital day 10, apixaban was held after administration of the morning dose in anticipation of a paracentesis, which removed 6.2 L of fluid without immediate complication on hospital day 11. Apixaban was not restarted after the procedure. On the morning of procedure, relevant labs included creatinine 4.4 mg/dL. A bilirubin drawn the following morning was 0.5 mg/dL and INR was 1.4 (had been 1.1 on admission). The morning after paracentesis the patient had a drop in hemoglobin from 7.9 to 5.5 g/dL and had worsening hypotension. An apixaban level by anti-Xa activity was measured at 50 ng/mL despite being more than 24 hours after the last dose. A CT angiogram 30 hours after paracentesis was suspicious for active extravasation along the left abdominal wall (Figure 1B), but interventional radiology angiogram was negative for active extravasation. The patient received resuscitation and a single dose of FEIBA for anticoagulant reversal. Although the patient achieved hemostasis, he continued to suffer multiple medical complications. He and his family elected to pursue palliative care 2 weeks after his hemorrhage, and he died shortly afterward.
Discussion
The risk of major hemorrhage associated with paracentesis has been estimated at less than 1%. 1 Even in patients with significant coagulopathy secondary to cirrhosis, the bleeding risk is low and the utility of correcting coagulopathy has been questioned. 2 Major professional societies, including the American College of Cardiology (ACC), 3 the American Association for the Study of Liver Disease (AASLD) 4 and the Society of Interventional Radiology (SIR) 5 classify paracentesis as a low bleeding-risk procedure. Both AASLD and SIR recommend against transfusing blood products to reverse liver disease-associated coagulopathy. While hemorrhage is a rare complication of paracentesis, 30-day mortality after hemorrhage may be higher than 40%. 6
Apixaban is a direct oral anticoagulant (DOAC) which directly inhibits factor Xa activity. In healthy patients, the maximum blood concentration occurs 3 to 4 hours after oral administration with a half-life of 12 hours. A recent systematic review suggested that apixaban had improved efficacy and similar safety compared to warfarin in the treatment of atrial fibrillation. 7 Elimination is a combination of metabolism as well as renal and hepatic excretion. 8
Patients with significant liver disease have been excluded from major trials of apixaban.9-11 While initial data have not demonstrated a significant increase in AUC (1.03 and 1.09) in Child-Pugh A and B liver disease, there is insufficient data to determine dosing recommendations in more advanced liver disease. 8 More recent data in patients with Child-Pugh B and C liver disease treated with edoxaban and apixaban showed an approximate 100% to 250% increase in Anti-Xa levels compared to patients with Child-Pugh A liver disease. 12 Labeling approved by the FDA recommends no dose adjustment of apixaban in patients with Child-Pugh A liver disease. 13 Due to coagulation abnormalities and limited data, the manufacturer cannot provide dosing recommendations in Child-Pugh B, and recommends against the use of apixaban for patients with Child-Pugh C. While no dosage adjustments are recommended in isolated renal impairment, the dose of apixaban prescribed for atrial fibrillation is reduced from 5 mg twice a day to 2.5 mg twice a day in patients with 2 of the following factors: age 80 years or older, body weight 60 kg or less, or serum creatinine 1.5 mg/dL or higher.
The manufacturer recommends holding apixaban for 48 hours in patients undergoing medium- and high-risk bleeding procedures and 24 hours for low-risk bleeding procedures. For major surgery in patients with atrial fibrillation, the PAUSE trial demonstrated that a brief interruption of 24 to 48 hours of apixaban was associated with an arterial thromboembolism rate of 0.16%. 14 More than 85% of patients holding apixaban for 24 hours had blood levels less than 50 ng/mL and the risk of bleeding was approximately 1%. While this suggests that brief periprocedural interruptions of apixaban are reasonable, these results were in a predominantly low thrombotic-risk population undergoing invasive surgery. The risks of thrombosis are likely higher in medically comorbid patients, including those with liver disease, and the intrinsic procedural risk of bleeding associated with paracentesis is likely significantly lower than in this population.
We found limited data on performing paracentesis in patients being actively treated with anticoagulation and no published data on the safety of DOACs in patients undergoing paracentesis. Neither platelet count nor INR have been shown to be predictive in hemorrhagic risk following paracentesis. 15 This observation may lead clinicians to assume that warfarin, which also raises the international normalized ratio (INR), may not increase paracentesis bleeding risk. However, within the published literature there is minimal evidence to directly support this hypothesis. A small study in 32 patients with Budd-Chiari Syndrome on warfarin with a mean INR of 3.1 and undergoing paracentesis for ascites found no increased risk of abdominal wall bleeding or hemoperitoneum. 16 However, apixaban and other DOACs inhibit coagulation through alternative mechanisms which may increase bleeding risk even if warfarin is safe to continue. An observational study of 14 patients on DOACs who received 87 large volume paracenteses reported only a single episode of minor bleeding. 12
While the strength of the association between apixaban and bleeding in our patients remains circumstantial, the occurrence of 2 major hemorrhage events after paracentesis performed by experienced providers under ultrasound guidance within a brief window of time is concerning for possible causation. While the patient in Case 2 had what was felt to be an adequate interruption in anticoagulation, his high residual apixaban blood levels suggest that a potentially therapeutic level of drug was still present in his body at the time of the procedure despite no elevation in total bilirubin. The Naranjo et al 17 adverse drug reaction algorithm score indicates a probable cause in both cases. A review of the FDA Adverse Events Reporting System Dashboard indicates multiple cases of apixaban-induced events with abdominal cavity drainage, with at least 2 cases classified as serious events. 18
Conclusions
We report major hemorrhages after routine paracentesis in 2 patients with moderate cirrhosis, moderate renal impairment, and chronic apixaban use. Despite a normal bilirubin level, the apixaban level remained high 24 hours after the last dose in one patient. In combined renal and hepatic impairment, laboratory parameters including bilirubin may not be reliable predictors of apixaban elimination. In the absence of high-quality evidence for the safety of continuing apixaban, we propose that apixaban should be interrupted for at least 48 hours in patients with cirrhosis undergoing planned paracentesis until better data from large observational cohorts is available. Liver failure patients (Child-Pugh B or C) may be at an increased risk for elevated apixaban blood levels. Providers may consider obtaining an apixaban anti-Xa level prior to paracentesis to assure level elevation resolution, but there is no consensus as to the safe level of apixaban to proceed. Multidisciplinary shared decision making should be utilized when deciding apixaban pre-procedure hold duration in patients with liver failure undergoing paracentesis.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
