Background and aims: To assess the risk of bleeding after percutaneous liver biopsy, we retrospectively analyzed 629 procedures with particular respect to patients with an increased a priori bleeding risk.
Methods: Factors possibly related to the risk of bleeding were analyzed by univariate analysis. Those variables which were significant in the univariate analysis were then entered into a forward conditional logistic regression model.
Results: Biopsy-related bleeding events defined as clinically overt complication (n = 10; 1.6%), an otherwise unexplained drop in serum hemoglobin concentration of greater than 2 g/dl (n = 45; 7.1%) or intra- or extrahepatic hematoma assessed by ultrasound (n = 17; 2.7%) were identified in 72 patients. 58% of the bleeding events occurred in patients with particular risk factors for bleeding. Biopsy-related mortality in the study cohort was 0.48%. Logistic regression analysis indicated mycobacterial infection [odds ratio (OR) 24.0], pre-biopsy prophylactic platelet substitution (OR 9.9), acute liver failure (OR 9.1), heparin administration on the day of biopsy (OR 8.7), advanced liver cirrhosis (OR 5.1), therapy with corticosteroids (OR 3.5) or metamizole (OR 2.8) and leukemia or lymphoma (OR 2.8) as significant (p < or = 0.05) independent risk factors. Delayed bleeding (>24 h after biopsy) was identified in 70% of the bleeding events.
Conclusions: In our study cohort which comprised a high proportion of patients with particular risk factors for bleeding, biopsy-related bleeding occurred more frequently and later than commonly observed and was associated with only a few prognostic factors. Considering these predictors before liver biopsy will aid to reduce the rate of bleeding complications.
Copyright 2003 S. Karger AG, Basel