ERCP and EUS Procedures: When Standard ERCP Fails

0:00Welcome to Board Pearls. This is episode four of seven of the ERCP and EUS Procedures chapter, in the Endoscopic Procedures module. In this episode we cover what to do when the standard route fails: the EUS-guided rescues for a duct you cannot cannulate, and the altered surgical anatomy that puts the papilla out of a duodenoscope's reach.

0:20Start with the moment the standard procedure fails. You have a patient with a dilated bile duct and you have tried to cannulate the papilla, and it will not happen. Tumor has distorted the medial duodenal wall, or the wire will not cross the stricture, or forty minutes of attempts have produced nothing but a frustrated endoscopist. The old answer was to stop, send the patient to interventional radiology for a percutaneous drain, and accept an external bag. The modern answer is that you already have the tool in your hand. The echoendoscope sees the dilated duct directly through the wall of the stomach or duodenum, and once you can see a dilated duct, you can puncture it. EUS-guided biliary drainage is the rescue, and the whole technique rests on one fact. A dilated duct is a target. If the duct is not dilated, you have nothing to aim at, which is why this rescue belongs to obstruction with upstream dilation and not to a normal-caliber system.

1:14So you puncture the duct under sonographic guidance, inject contrast to get a fluoroscopic roadmap, and now you face the real decision. There are three ways to finish, and the anatomy chooses for you. The single question that organizes all three is this. After I have a wire in the duct, where can it go, and what does that let me preserve.

1:34The first path is the one you want whenever you can have it. EUS-guided rendezvous. You puncture the duct, you advance the wire antegrade, down through the stricture, across the papilla, and out into the duodenal lumen. Then you withdraw the echoendoscope and exchange for a duodenoscope. You grab the wire poking through the papilla and complete an ordinary ERCP over it. Think about what that accomplishes. You have not created any permanent opening. Bile still drains the natural way, through the papilla, into the duodenum, the way it always did. You used the EUS puncture only as a temporary detour to get a wire where your retrograde cannulation could not. The tract you made through the duct wall closes on its own. Rendezvous is the lowest-burden rescue because it leaves the anatomy intact, which is exactly why it sits at the top of the hierarchy. The one requirement is that the papilla has to be reachable, because the finishing move is a normal duodenoscope coming up to retrieve the wire. If you cannot get a scope to the papilla, rendezvous is off the table no matter how perfect the puncture.

2:38So picture the case where the papilla cannot be reached. A pancreatic head tumor has obstructed the duodenum, or a duodenal mass has invaded the medial wall, or the surgical anatomy simply does not let a scope get there. The wire can go into the duct, but it cannot exit through a papilla you can never retrieve it from. Now you ask the second question. Is the extrahepatic duct accessible from the duodenal bulb. Say the dilated common bile duct sits right against the bulb, within a centimeter, with no vessels in the way on Doppler. Then you go directly. EUS-guided choledochoduodenostomy. You create a fistula straight from the duodenal bulb into the extrahepatic duct and you hold it open with a lumen-apposing metal stent. The mechanism is the whole point of the device. A LAMS has two flanges, one on each side, and they pull the duodenal wall and the duct wall together and clamp them. That apposition is what lets the new fistula mature without bile leaking into the peritoneum while the walls fuse. Bile then drains through the stent into the duodenum, bypassing the obstructed papilla entirely. This is the standard rescue for a distal malignant obstruction when the papilla is unreachable but the extrahepatic duct is right there and intact.

3:52Then there is the hardest case, where neither the papilla nor the extrahepatic duct is available to you. Hilar obstruction isolating an intrahepatic territory. A prior Whipple. A gastric bypass. The extrahepatic duct is encased by tumor or is simply out of reach, so you cannot drain from the bulb. What you can still see, from the proximal stomach, is a dilated left intrahepatic duct. So you puncture that, from the stomach, and you run a stent from the gastric wall into the intrahepatic biliary tree. EUS-guided hepaticogastrostomy. The stent here is built differently for a reason. It is partially covered. The covered segment sits in the tract between stomach and liver, and the membrane seals that tract so bile does not weep along it. The uncovered segment sits up inside the intrahepatic duct, and leaving it bare matters, because a covered membrane there would wall off the side branches you are trying to drain. So the design follows the function precisely. Seal the tract, leave the intraductal portion open.

4:50Now hold the three side by side, because the selection logic is the teaching point and the boards test it as a selection problem. Rendezvous when the papilla is reachable. Choledochoduodenostomy when the papilla is not reachable but the extrahepatic duct is accessible from the bulb. Hepaticogastrostomy when neither is available and only a dilated left intrahepatic duct remains. The ranking is not arbitrary. It tracks how much natural anatomy you get to keep. Rendezvous preserves the papilla and makes no permanent fistula at all. Choledochoduodenostomy makes one short fistula into the duct that was closest. Hepaticogastrostomy makes the deepest fistula, stomach to liver, and so it is the last resort. The organizing instinct is to preserve natural drainage and to make the smallest possible new opening that the anatomy permits. Operator expertise and whether the dedicated stents are stocked also weigh in, but the anatomic hierarchy is what drives the choice.

5:48Hepaticogastrostomy carries a complication story that is worth getting exactly right, because the framing has shifted and the old framing is a trap. The historical teaching was that the dominant complication of hepaticogastrostomy is stent migration into the peritoneum. The reasoning was sound. The gastric wall is mobile, it moves with every breath and every meal, and the anchoring length up in the intrahepatic duct is short. A short anchor in a moving wall predicts a stent that works itself loose. So engineers responded. Dedicated hepaticogastrostomy stents now combine that covered tract segment with proximal flares and intraductal flanges and anchoring flaps, all designed to resist the stent sliding along its axis. And those designs worked. With the dedicated anti-migration stents, migration has become uncommon. So if a question hands you a hepaticogastrostomy patient and you reflexively say the most likely problem is migration, you may be answering for the era before the device improved. With current stents, the events you actually see most are bleeding, bile leak, and cholangitis. The cholangitis has its own mechanism worth holding. You have made a permanent open channel from the stomach into the biliary tree, so gut flora can ascend through the stoma and colonize the bile. Picture the patient who comes back a few days after hepaticogastrostomy. There is fever, right upper quadrant pain, rising bilirubin, and a fluid collection around the tract. That is the leak-and-cholangitis picture, not the migrated-stent picture, and the stent is sitting right where you left it. Do not call migration the dominant complication categorically, because the corrected framing is that anti-migration design moved it down the list.

7:30Now turn from the duct you cannot cannulate to the papilla you cannot even reach, because the surgeon got there first. This is ERCP in altered anatomy, and it runs on a single principle that makes every choice fall out. The side-viewing duodenoscope wants an en face view of the papilla, looking straight at it, because that orientation is what lets you cannulate and cut along the duct's axis. Surgical anatomy puts barriers between the scope and that view, and the technique you choose is whichever one restores en face access through the particular barrier in front of you. Three anatomies come up, and each barrier is different.

8:08Roux-en-Y gastric bypass first, because it is the one that breaks standard ERCP most completely. The surgeon made a small gastric pouch that drains into a Roux limb, and that limb bypasses the duodenum entirely. The old stomach, the duodenum, and the papilla are all sitting in the excluded segment, off the main food stream. To reach the papilla the natural way you would have to go down the Roux limb, all the way to the jejunojejunostomy, then turn and climb back up the biliopancreatic limb to the duodenum. That is a hundred to a hundred fifty centimeters or more. A standard duodenoscope is nowhere near long enough, so conventional ERCP is essentially impossible in this patient, and that impossibility is the whole reason specialized approaches exist.

8:52The modern preferred answer is the EDGE procedure, EUS-directed transgastric ERCP, and it is a clever use of the same fistula logic from the rescue section. You take the EUS scope and you find the excluded stomach, which usually sits right next to the gastric pouch or the Roux limb across a thin tissue plane. You deploy a lumen-apposing metal stent between them, creating a temporary fistula. Note that the fistula can run two ways. It can be gastrogastric, from the pouch to the old stomach, or jejunogastric, from the Roux limb to the old stomach. Which one depends on the access route. Either way, once that fistula matures, you have built a shortcut. You pass a standard duodenoscope through the LAMS into the excluded stomach, then down through its pylorus into the duodenum. Now you are looking at the papilla en face, exactly as you would in a patient who never had surgery. That is the technical principle. EDGE recreates the normal retrograde route to the papilla by tunneling into the bypassed stomach. No surgical incision, standard duodenoscope, and you can do several ERCP sessions through the same fistula if you need to.

9:59But the fistula is the catch as well as the trick. A persistent communication from the pouch back into the old stomach undoes the restriction the bypass was built to create, and the patient regains weight. So the rule is that at the end of biliary therapy you remove the LAMS and close the fistula. A patient who returns weeks later, stones cleared and liver enzymes normal, worried about weight regain, gets the fistula addressed precisely for that reason. You do not leave a gastrogastric channel open just because it was convenient.

10:28The alternative to EDGE in bypass anatomy is balloon enteroscopy, single-balloon or double-balloon, threaded all the way down the Roux limb to reach the papilla from below. It can be done, but the therapeutic success runs only around sixty to seventy percent, and the reason is the principle we started with. An enteroscope is forward-viewing. It looks straight ahead, not laterally, so it cannot give you the en face papillary view a side-viewing duodenoscope provides. You are trying to cannulate a target you are seeing edge-on instead of face-on, and the cannulation and therapy suffer for it. That gap is exactly the access EDGE restores, which is why EDGE is preferred where the expertise exists. There is also an older laparoscopy-assisted approach. A surgeon creates a gastrostomy in the remnant so a duodenoscope can go in directly, which gives true standard access. But it costs you general anesthesia, surgical involvement, and a gastrostomy, so current guidance favors EDGE over it.

11:26The second altered anatomy is Roux-en-Y hepaticojejunostomy, after a Whipple or a benign biliary reconstruction. Here the surgeon sewed the bile duct directly onto a Roux limb of jejunum and bypassed the papilla altogether. So there is no papilla to cannulate. Your target is the anastomosis itself, and the access problem is the same long Roux limb. You reach the hepaticojejunostomy with balloon-assisted enteroscopy, single or double balloon. Through it you can dilate an anastomotic stricture, extract stones, or place a stent. Success here runs higher than at the papilla, around sixty to eighty percent, partly because dilating an anastomosis does not demand the same precise en face cannulation that the native papilla does. When enteroscopic access fails, percutaneous transhepatic cholangiography is the fallback. EUS-guided hepaticogastrostomy, the same technique from the rescue section, is the modern alternative when you have a dilated left intrahepatic duct to puncture.

12:27The third anatomy is Billroth two, the older gastrectomy you still see in elderly patients. The surgeon made a gastrojejunostomy with two limbs. The efferent limb carries food away from the stomach remnant, and the afferent limb carries bile and pancreatic juice up from the duodenal stump. To reach the papilla you have to identify the afferent limb and travel up it, retrograde, to the duodenal stump where the papilla lives. A standard duodenoscope can sometimes make that trip, but the perforation risk is markedly elevated, and the numbers are the corrected teaching point. Sphincterotomy-related perforation runs roughly one and a half to five percent. And overall perforation has been reported as high as five to eight percent. That number counts injury to the afferent limb and the scope-passage trauma of forcing a side-viewing scope retrograde up a limb it was not designed to navigate. Compare that to well under one percent for standard ERCP. So the danger in Billroth two is not only at the cut, it is in the journey. That is why a forward-viewing endoscope is often preferred for the afferent limb intubation, even though you give up the side-viewing papillary view to use it. A cap on the scope tip helps you see and cannulate.

13:36And there is one more inversion in Billroth two that the boards test directly. Coming up the afferent limb, you approach the papilla from the opposite side, so the papilla is oriented upside down relative to standard ERCP. That changes the cut. The safe direction for the sphincterotomy is reversed, opposite to the direction you would cut in normal anatomy. The reason is that you aim the incision along the duct axis, and the axis now points the other way. Cut the standard direction here and you cut away from the duct and toward perforation. Controlling the cut direction is harder in this orientation. So a needle-knife sphincterotomy made over a previously placed stent is often safer than a standard pull-type sphincterotome, because the stent gives you a fixed track to cut along. The mechanism is the same one running through the whole episode. The orientation of the papilla dictates the safe direction of the cut, and the surgical anatomy dictates the orientation.

14:35So pull the way of thinking together. When the standard route fails, you do not abandon endoscopy, you ask one question about anatomy and let the answer pick the technique. For a duct you cannot cannulate, the question is where the wire can go after you puncture under EUS. The answer ranks rendezvous over choledochoduodenostomy over hepaticogastrostomy, preserving as much natural drainage and making as small a fistula as the anatomy allows.

15:02For a papilla the surgery has hidden, the question is how to restore the duodenoscope's en face view through the particular barrier. The answer is EDGE for the bypass, balloon enteroscopy for the hepaticojejunostomy, and a reversed, cautious approach up the afferent limb for Billroth two. In both halves the device design and the technique are not memorized lists. They are consequences of the anatomy, and if you can name the barrier you can almost always reconstruct the rescue.

15:31The next two episodes move from access problems to the pancreas itself, covering EUS and ERCP across the pancreatitis spectrum. The first takes the gland, the workup of recurrent acute pancreatitis and the decompression of the chronic dilated duct. The second takes the collections, the step-up drainage of pancreatic fluid collections and direct endoscopic necrosectomy. The organizing question throughout is when the modest benefit of intervention is worth the very real risk of the procedure.

16:09For the full chapter, the practice vignettes, and the topic-tagged question bank, head to board pearls dot com. You'll find the rest of the series on Apple Podcasts, Spotify, or wherever you listen to podcasts. That brings us to the end of episode four of seven of chapter twenty nine, and I'll see you in the next one.