Background: Initial experience suggests that EUS-PBD holds promise. Aims & Methods: To assess the feasibility, technical requirements, safety and efficacy of EUS-BPD in pts with previously failed standard ERCP drainage. Between Nov03-Nov06, 58 pts out of 2514 consecutive ERCPs had failed biliary or pancreatic drainage (2.3%). The indication was revised in 17 cases and EUS-BPD considered for the remaining 41 (male/female: 26/15; age range: 36-91 yr). Indications, drainage techniques, technical & clinical success and complications were reviewed. Therapeutic EUS-scopes (Pentax 38-UX in 27 & Olympus GF-UCT160 in 14), EUS-FNA needles of 19 or 22 G (30/8) with .035" & .018" guidewires, and fluoroscopy were used. Trends over the 3-yr period were also analyzed. Results: Pancreatic ductal drainage was attempted in 4/41 (2 Whipple's, 2 chronic pancreatitis) and biliary drainage in 37 (12 malignant papillary involvement,11 contralateral duct access problems in hilar strictures, 8 prior Roux-en-Y & 6 miscellaneous), all but 4 with malignant obstructive jaundice (20 primary, 13 metastatic, 4 unknown). In 3/41 ductal puncture was precluded by lack of an obvious target, interposed vessels, ascites and/or scope instability. Ductography was obtained in 37/38 punctures (1 parenchymal injection), ductal guidewire access in 34, and stent placement in 30, 16 metal & 14 7F plastic stents. Thus, EUS-BPD was technically successful in 30/38 punctures (79%), 21 transmural (through the liver/bile-duct/pancreas = 16/4/1) & 9 transpapillary (rendezvous/antegrade = 6/3). Clinical success (resolution of jaundice or pain) was achieved in 22/30 (63%) and failed in 2, whereas the outcome could not be assessed in 6 pts with early deaths not procedure-related. There were 7/38 (18.4%) complications, 4 mild (bile leaks that settled in 3-7 days), and 3 severe (pseudocyst, biloma & hemoperitoneum, 1 death). Cumulative technical success/complications rates at 1,2 & 3 yrs were in% 50/38, 73/31 & 79/18.4 (p < 0.05). Qualitative aspects judged relevant for improved outcomes include: Video Vs Fiber EUS scopes, 19 Vs 22G needles, cautery Vs dilators for access, OTW protrected Vs freehand stent exchange. Conclusions: EUS-PBD is a demanding procedure, with a success rate of around 80%, and with a morbidity close to 20%. These results are less encouraging than previously reported. However they reflect a learning curve effect and initial equipment & device selection limitations with a strong time trend to be overcome. EUS-PBD currently offers unique advantages for selected patients, and may soon realize the expected advantages over PTC for the vast majority of patients with failed ERCP drainage.