A total of 1192 consecutive patients with diffuse liver disease were randomized to have percutaneous liver biopsy specimens taken with the Menghini or the Tru-Cut needle, to compare tissue yield, safety, and accuracy of the two needles for diagnosing cirrhosis. The sites of puncture were determined by prebiopsy ultrasound scans. Adequate samples were obtained from 94% with the Tru-Cut needle and from 79.2% with the Menghini needle (p less than 0.001). Accuracy in diagnosing cirrhosis was 89.5% for the Tru-Cut needle and 65.5% for the Menghini needle (p less than 0.05). Complication rates were very low and similar for both needles. Under these conditions, the Tru-Cut needle is superior to the Menghini needle for diagnosing cirrhosis.