Preoperative administration of the simple bile salt sodium deoxycholate has been shown in this study to prevent postoperative endotoxaemia and renal failure in patients with obstructive jaundice. Fifty-four per cent of jaundiced patients not given the salt were found to have systemic endotoxaemia, associated with renal impairment in two-thirds of the cases. No patient given sodium deoxycholate 500 mg 8 hourly for 48 hours before operation had portal or systemic endotoxaemia, and none had evidence of renal impairment (P less than 0 X 02, X2 with Yates' correction). The incidence of endotoxaemia in untreated jaundiced patients was very significantly greater than in non-jaundiced patients undergoing elective upper abdominal surgery (P less than 0 X 005), but this difference is abolished by the prophylactic administration of the oral bile salt. The mechanism of action of bile salts in preventing endotoxin absorption from the small bowel has been investigated, and the lack of any significant alteration in the small bowel microbial flora in obstructive jaundice suggests that a direct effect on the endotoxin molecule is involved. Nearly 20 per cent of patients with obstructive jaundice still develop postoperative renal insufficiency, but preoperative prophylactic use of sodium deoxycholate should reduce this very significantly.