RISK OF PERCUTANEOUS ENDOSCOPIC GASTROSTOMY-RELATED BLEEDING WITH UNINTERRUPTED ANTICOAGULATION: A MULTI-CENTER RETROSPECTIVE COHORT STUDY

Introduction: Percutaneous endoscopic gastrostomy tube placement is often essential for patients with cerebrovascular diseases related to thromboembolic events, leading to neurologic dysphagia. These patients are commonly maintained on anticoagulant treatment to prevent recurrence of thromboembolic events. Also, PEG is indicated in patients with malignant luminal obstruction and those patients are frequently on anticoagulation given the high rate of venous thromboembolism associated with malignancy. This study aimed to assess bleeding risk associated with uninterrupted anticoagulants preceding PEG placement through a multi-center retrospective cohort analysis.

Methods: Utilizing the TriNetX US Collaborative Network, we identified patients who underwent PEG procedures while on anticogulation and compared them to a control group of patients who didn’t receive any anticoagulant therapy prior to the procedure. Anticoagulant therapy included apixaban, edoxaban, rivaroxaban, dabigatran, and warfarin. The primary outcome was gastrostomy-related bleeding within 30 days, with secondary outcomes including all-cause mortality, ICU admission, and the need for blood transfusion during the same period.

Results: Our cohort comprised 2802 patients in the anticoagulants (AC) group and 66880 in the control group, with a mean age of 68 +/- 14 years in the AC group and 59.6 +/- 18.6 in the control group. Both groups included a higher percentage of males, Whites, and non-Hispanic/Latino patients. Additional baseline characteristics are detailed in Table 1. Propensity score matching balanced age, sex, race, ethnicity, anti-platelet medications, and comorbidities (e.g., diabetes, hypertension, cerebrovascular diseases, obesity, and chronic kidney disease) (Table 1). After matching, the rate of gastrostomy-related bleeding within 30 days was 1.07% in the AC group and 0.57% in the control group (P value 0.038). No statistically significant increase was observed in blood transfusion or ICU admission rates in the AC group compared to the control group (Table 2).

Conclusion: In this US-based population study, patients receiving anticoagulation displayed a higher risk of gastrostomy-related bleeding compared to the control group. However, the absolute increase in bleeding risk was small (Risk difference 0.5%). Furthermore, rates of severe bleeding, as indicated by ICU admission and blood transfusion, was not statistically different beween the two cohorts. These findings aim to inform the clinical decision-making in pre-procedure preparation of Percutaneous Endoscopic Gastrostomy.