The authors reviewed the clinical charts and the radiographic files of 93 patients with obstructive jaundice--in 86 cases due to neoplasm--treated with PTBD. The test of differences from survival curves was used to identify the clinical parameters predictive of short survival after PTBD. The difference in survival curves was significant relative to serum indirect bilirubin (cut point: 7.6 mg%), to serum cholinesterase (cut point: 1290 mU/ml), to white blood cell counts (cut point: 8600/mm3), to blood urea nitrogen (BUN) levels (cut point: 60 mg%). Because of the marked negative prognostic value of high BUN levels, our data seem to indicate that PTBD should not be performed when severe renal insufficiency is present. Other parameters correlated with a short survival after PTBD were the histotype of metastasis (in comparison with the other ones), and large neoplastic volume (in comparison with a small and a medium ones). Through pre-PTBD radiological and laboratory data analysis, a group of patients can be selected in whom the procedure will increase neither well-being nor survival, as plotted against those patients who are likely to benefit from biliary drainage.