The usual technique of collecting gallbladder bile at laparotomy is not suitable for sequential studies of cholesterol nucleation time (NT) in patients receiving therapy to prevent or dissolve cholesterol gallstones. Our aim was to study the feasibility of measuring NT in bile obtained by nasobiliary or nasoduodenal intubation. We studied a total of 10 cholesterol gallstone patients; in 8 bile was collected by nasobiliary drainage, in 7 it was collected by nasoduodenal intubation, and in 3 it was collected at laparotomy the next day. Three patients developed abdominal pain and increased serum amylase after endoscopic retrograde cannulation. All three biles obtained at operation nucleated quickly (NT, 1-4 days), whereas duodenal biles were all beyond the expected range (NT, > 21 days). Chymotrypsin activity, as a marker of pancreatic juice contamination, was detected in five of eight nasobiliary biles and in all seven duodenal biles but in none of the surgical biles. Free fatty acids (reflecting lipolysis) were significantly higher in duodenal than in surgical biles, with nasobiliary bile showing intermediate values. Nasobiliary bile showed either a rapid (median NT, 3 days) or a slow (median NT, 22 days) NT, depending on whether chymotrypsin activity was absent or present (p < 0.05). We conclude that duodenal bile is never suitable for NT determination because of contamination by pancreatic enzymes, and that nasobiliary bile, if not contaminated by pancreatic enzymes, may be suitable for NT determination but that its collection via a nasobiliary tube after cholecystokinin injection carries a risk of pancreatitis.