Patients with liver disease frequently have impaired blood coagulation. The optimal method for liver biopsy in this situation is not established. To investigate this issue we randomised 117 patients with impaired blood coagulation, in whom liver biopsy was required, to receive either transjugular or plugged-percutaneous biopsy. Seventeen patients were excluded prior to biopsy and a protocol biopsy was performed in 100 patients (44 transjugular, 56 plugged-percutaneous). Liver tissue was obtained in 97 (42 transjugular, 55 plugged-percutaneous). Plugged-percutaneous liver biopsy was quicker and easier than transjugular liver biopsy and the biopsies obtained were significantly larger (12 +/- 5 mm vs. 6 +/- 4 mm; p < 0.001). However, 2 of 56 (3.5%) patients who received plugged-percutaneous biopsy had haemorrhage which required transfusion, while none of the 44 patients who received transjugular biopsy had haemorrhage (not significant). Both methods of liver biopsy were associated with a high success rate and a low incidence of complications. Plugged-percutaneous liver biopsy provides larger biopsies but may be associated with an increased risk of haemorrhage.