A transjugular liver biopsy was performed on 60 patients. Specimens were successfully obtained from 57 (95%) patients. Specimens obtained from cirrhotic patients were frequently small-sized/fragmented. The wedge hepatic venous pressure and hepatic venous pressure gradient were higher in patients with small-sized/fragmented specimens than those with non-fragmented specimens (16.3 +/- 6.4 vs 12.3 +/- 4.9 and 10.9 +/- 6.2 vs 7.3 +/- 3.4 mmHg, P < 0.05, respectively). During the same period of time, percutaneous liver biopsies were consecutively performed on 277 patients. The liver specimens by transjugular method were generally smaller (0.63 +/- 0.58 vs 1.50 +/- 0.86 cm, P < 0.001) and more fragmented (63% vs 16%, P < 0.01) than those obtained by percutaneous method. Biopsy specimens obtained for diagnosis by the former method were inadequate from 6 (10%) patients and by the latter route were inadequate from 7 (2%) patients. Subcapsular haematoma in one patient was associated with the transjugular liver biopsy. Minor complications occurred in three patients: neck haematoma in two and paroxysmal supraventricular tachycardia during the procedure in one. In comparison, percutaneous liver biopsy was followed by minor complications in 20 patients and major complications in four patients. It is concluded that transjugular liver biopsy is a safe, valuable and alternative procedure to obtain liver specimens, especially in patients who were contraindicated for percutaneous liver biopsy.