EUS pancreatic rendezvous to rescue failed ERCP for postoperative pancreatic fistulas after pancreatic or duodenectomy. These are my disclosures. Pancreatic fistulas after Whipple can be classified as clinically insignificant, grade A, or clinically significant in which grade B has a duration of greater than three weeks and requires percutaneous or endoscopic drainage, and grade C, which in addition is associated with organ failure. Clinically significant pancreatic fistulas occur in 25% to 30% of patients and are associated with disease-specific risk factors such as smaller pancreatic ducts and softer glands, patient-related factors such as age, obesity, malnutrition, and the surgical techniques such as the type of anastomosis and blood loss. They are diagnosed with drain amylase levels greater than three times the serum amylase. Management involves supportive nutrition when terral nourishment is as good as TPN. Octreotide may have mild benefit, but the mainstay of treatment is drainage, most commonly done percutaneously. Median duration of drains for grade B is 30 days and 40 days for grade C fistulas. We present three cases of grade B and C fistulas who failed percutaneous drainage alone and conventional ERCP but were rescued with EOS pancreatic rendezvous. Their median age was 62 and median BMI was 40. They all had a soft gland and the median pancreatic duct preoperatively was only 2.5 millimeters. Also, two of the three patients had intraoperative stents placed. Case 1. A 56-year-old male, 28 days post-whipple, despite two indwelling percutaneous drains, was admitted with wound dehiscence and pancreatic juice leaking from his incisions and sepsis. A CT scan showed fluid tracking alongside the drain and through the ventral surgical incision despite an adequately placed drain near the pancreatic or jejunostomy. So the patient is taken for an ERCP to better control this leak. Frank passes C in emerging as attempts are made to cannulate the pancreatic duct, with contrast injection and via passage showing no discernible duct. On EOS, a 1.5-millimeter pancreatic duct is identified. The first attempt at a 19-gauge needle puncture and contrast injection is unsuccessful. The second pass, however, obtains a pancreatogram, but no contrast flows into the jejunum. A 0.025-inch angled guide wire is passed carefully, gradually pulling the needle back till the duct is ultimately cannulated with the wire. But the wire needs to be passed towards the pancreatic or jejunostomy, so with a steady needle tip barely in the duct, the wire is manipulated till it flips downstream. Now the wire coils in the dehisced space between the pancreas and jejunum. With continued and careful manipulation, ultimately we find the pancreatic or jejunostomy, and the wire is passed into the jejunum. Note, the dehiscence is at least 3, if not 4 centimeters long. The scope is then withdrawn, leaving the wire in the pancreatic duct. An exchange for a pediatric colonoscope, which is carefully advanced to the pancreatic limb without dislodging the wire. The guide wire is finally seen. Unable to snare it in this position, the wire is grasped with a forcep and pulled down to a more stable position to allow snaring it to rendezvous. The wire is carefully pulled through the scope while advancing the wire through the mouth. A 9 to 12 extraction balloon is then advanced over this rendezvoused wire, and a pancreatogram obtained. The rendezvoused guide wire is then completely withdrawn, an exchange for a new 0.025 inch guide wire passed into the tail of the pancreas, over which a 7 French by 9 centimeter pancreatic duct stent is placed. The intraoperatively placed but migrated pancreatic duct stent is removed. Post ERCP day 8, he develops another fever, and CT scan shows a fluid collection closer to the tail of the pancreas consistent with an abscess. This is easily drained with a 10 by 10 millimeter cotri-enhanced luminoposing metal stent, which is then removed after abscess resolution three weeks later. Post ERCP day 72, he is taken for an ERCP to remove his stent, but there's still a space between the pancreas and the jejunum, and again a failure to cannulate the pancreatic duct, even though he is asymptomatic at this time. So he undergoes a second EOS pancreatic rendezvous, allowing placement of two 7 French pancreatic duct stents, which after a dwell time of six months are removed at an ERCP. A patent pancreatic or jejunostomy is seen. A 0.025 inch guide wire is passed into the pancreatic duct, and a pancreatogram shows no more leak. He lived for an additional 33 months with no recurrent leaks or pancreatitis till he succumbed to his metastatic cancer. For the three patients treated with EOS pancreatic rendezvous who had failed ERCP, they had a median of two drains for a median duration of 30 days prior to the EOS. Median pancreatic duct size on EOS was small at two millimeters, and we used one or two 7 French stents in all cases. Percutaneous drains were removed after a median of 20 days after the EOS, and the median stent dwell time was 210 days. Median follow-up was 33 months with no adverse events or recurrences. Postoperative pancreatic fistulas after a Whipple that are clinically severe, grades B and C, may sometimes fail management with percutaneous drains alone. Conventional ERCP may be difficult if there is a large dehiscence. EOS pancreatic rendezvous can help bridge these leaks, but can be challenging due to the decompressed pancreatic duct. Careful needle puncture and wire manipulation can allow access and reconnecting the disrupted duct to the jejunum. Median time to drain removal may be shortened in these situations. In conclusion, when percutaneous drains and conventional ERCP fail to bridge a severe postoperative pancreatic fistula after a Whipple, don't give up. An EOS pancreatic rendezvous can help reconnect the disrupted duct to the jejunum and possibly hasten recovery. Further studies are needed to compare this approach to percutaneous drainage alone.

EOS pancreatic rendezvous

postoperative pancreatic fistulas

ERCP

Whipple procedure

pancreatic duct