Editor,—I read the recent occasional viewpoint by Shah et al ((1999) Gut 45:628–629.OpenUrlCrossRefPubMedWeb of Science) with much interest. The authors describe their regional practice of mandatory real time ultrasound guidance for percutaneous liver biopsy in all cases of suspected diffuse liver disease. However, the published literature does not convincingly support the universal adoption of such a policy.

The only prospective randomised study cited in support of their protocol, by Lindor and colleagues,1 is open to a number of methodological criticisms. In particular, in an unspecified proportion of the patents randomised to ultrasound guidance, the procedure was not actually performed under direct guidance and was instead immediately preceded by a “biopsy room” ultrasound scan. The net result of this may have been to selectively raise the pre-biopsy scanning rates in the “ultrasound” cohort who were already more likely to have been previously scanned than the “blind” patients (76–78% v 67–68% in the respective groups).

In terms of the reduction in post-biopsy complications claimed by Lindor et al, the major impact was a reduction in hospitalisation due to pain. There was no statistically significant difference in the rates of bleeding or hypotension. The reduction in …