Background: Medical providers are often hesitant to perform bedside procedures in patients with an INR >2 or in patients who have received chemoprophylaxis or therapeutic anticoagulation, despite previously conducted studies suggesting cirrhotic patients with elevated INRs are not at higher risk for complications from a paracentesis or thoracentesis (1)(2). The Hospital Medicine Advanced Practice Provider Procedure Team at Parkland Memorial Hospital (PMH), a 900-bed academic county hospital, performs about 2000 bedside procedures annually —including paracenteses, thoracenteses and lumbar punctures— in order to expedite workup and therapeutics in hospitalized patients. After changing our procedure team protocols to proceed with bedside paracenteses and thoracenteses irrespective of INR levels, chemoprophylaxis or anticoagulation status, we aimed to show that there is no increase in the rate of major bleeding in patients with an elevated INR or recent anticoagulant use.
Methods: A retrospective study was performed comparing the rate of major bleeding events between patients with INR < 2 versus INR >= 2 undergoing bedside paracenteses and thoracenteses on the hospital medicine service at PMH. A chart review was performed comparing 275 bedside procedural instances in patients with INR >= 2 to 272 instances when INR < 2 from February 2020 to August 2021, including patients receiving prophylactic and therapeutic anticoagulation. There were several patients receiving repeat procedures in each subgroup. The primary outcome was major bleeding events, defined as a >= 2g/dL drop in hemoglobin or one that required >=2 units of packed red blood cells based upon International Society of Thrombosis and Hemostasis guidelines (3). A test for difference in proportions was used to compare the two groups, and p-values of < 0.05 were considered significant.
Results: There were no significant differences in age or gender between the two groups (see Table 1). Fewer procedures in the INR < 2 subgroup involved patients with alcoholic hepatitis or alcoholic cirrhosis as compared with the INR >=2 subgroup (44.5% vs 75.6%, p<.01). More procedures in the INR < 2 subgroup (14.7% vs 5.8%, p<.01) were associated with non-HCC malignancy. More patients were on chemoprophylaxis (10.3% vs 4%, p<.01) and therapeutic anticoagulation (8.5% vs 2.9%, p=0.05) peri-procedure in the INR < 2 subgroup. All major bleeding events occurred after paracenteses. There was no significant difference in the incidence of major bleeding for INR < 2 vs INR >=2 (0.7% vs 1.8 %, p=.26). None of these patients received prophylactic or therapeutic anticoagulation peri-procedure. Of the 2 patients with INR < 2 with major bleeding, one developed hemoperitoneum, but neither required intervention by Radiology. Of the 5 patients with INR >= 2 with major bleeding, one did not require transfusion and two developed hemoperitoneum, of which one required embolization of the inferior epigastric artery by Radiology. No procedure resulted in mortality.
Conclusions: This study suggests that INR and anticoagulation do not significantly increase major bleeding events in bedside paracenteses and thoracenteses. It also suggests that INR or anticoagulation should not delay or preclude a provider from performing these procedures when necessary to guide management or provide therapeutic relief for similar patient populations. This study is limited by sample size. Future directions include expanding sample size and exploring other contributing risk factors.