Nineteen patients who received chenodeoxycholic acid 750 mg/day for six months had duodenal bile aspirated before and after treatment. In five patients with hypertriglyceridaemia but no gallstones cholesterol saturation was reversed in every case, the mean cholesterol saturation index (SI +/- standard deviation) changing from 1-38 +/- 0-31 to 0-68 +/- 0-06 (P less than 0-005). In 14 patients with gallstones there was also an improvement in bile cholesterol content, but this was not sufficient to produce mean unsaturation, saturation index changing from 1-55 +/- 0-52 to 1-13 +/- 0-43 (P less than 0-05). Only seven of 14 patients with gallstone achieved cholesterol unsaturation. In four patients with hypertriglyceridaemia and gallstones, mean unsaturation was produced and the saturation index changed from 1-70 +/- 0-45 to 0-86 +/- 0-47 (P less than 0-05). When all nine patients with hypertriglyceridaemia were grouped, the mean saturation index fell from 1-52 +/- 0-40 to 0-76 +/- 0-30 after therapy (P less than 0-001). In contrast the 10 patients without hypertriglyceridaemia showed no significant fall in saturation index which was 1-50 +/- 0-54 before and 1-24 +/- 0-40 after therapy. The ability of chenodeoxycholic acid feeding to improve bile saturation with cholesterol correlated with the presence of hypertriglyceridaemia whether or not gallstones were present. It did not correlate with gallstone dissolution or body weight.