Aspects on the role of prophylactic procedures to influence post-ERCP complication rates

Abstract

Background: When the technique to use ERCP was introduced almost fifty years ago, the morbidity in treatment of hepato-biliary diseases decreased due to the introduction of this miniinvasive modality, reducing the need for open surgical procedures. However, ERCP procedures are still marred with complications such as pancreatitis, cholangitis, hemorrhage and perforation and every measure must be undertaken to reduce these adverse events.

Objectives: The hypotheses of this thesis were: 1) Prophylactic antibiotics in ERCP do not reduce the complication rates enough to recommend it generally. 2) Prophylactic pancreatic stents reduce the PEP risk more the larger they are. 3) A grading scale for the complexity of the ERCP procedure (HOUSE) was validated in relation to success-rates, complications and duration of the procedure. 4) Preoperative SEMS in periampullary tumors show less bacterial contamination in intraoperatively collected bile than plastic stents, thereby reducing perioperative complications.

Methods: In the first study all ERCPs, included in GallRiks between May 2005 and June 2013, were studied regarding complication rates in relation to prophylactic antibiotics. Further, in the second paper, all ERCPs between 2006 and 2014 where an accidental pancreatic cannulation occurred and a prophylactic pancreatic stent was used were investigated, determining how the diameter and length of the stent affected the adverse events. In the third study, an ERCP complexity classification, (HOUSE), was validated in relation to success-rates, complications and duration of the procedure. The final study, an RCT compared preoperative SEMS to plastic stents in resectable periampullary tumors regarding intraoperative bacterial, histopathological and surgical technical findings as well as perioperative complications.

Results: In the first study complications were studied in relation to prophylactic antibiotics. We found a reduction of 26 % of OR in overall complications if prophylactic antibiotics were given, but in absolute figures reduction of the risk was a modest 2.6% and the NNT 38 patients to avoid one complication. In our second study an almost fourfold OR elevation (OR 3.58) in complication rates was seen if prophylactic pancreatic stents with a diameter ≤5 Fr were used compared to stents >5 Fr, the complication rates were further lowered (1.4 %) if the stents were >5 cm. The third paper validated a new three-graded ERCP complexity grading scale (HOUSE) in relation to success and complication rates, demonstrating a doubled PEP rate in HOUSE 2 and 3 (7.0 % and 6.8 %) compared to class 1 (3.4%) and longer procedure times, the higher the HOUSE class (HOUSE 1, 40 min; 2, 65 min; and 3, 106 min). In the final study, comparing preoperative SEMS to plastic stents in resectable periampullary tumors, higher preoperative stent dysfunction rates were found among the plastic stents (19 % vs 0 %, p=0.03). Intraoperatively, no differences were seen in bacterial occurrence in collected bile or in operative technical difficulties, but a higher histopathological foreign body reaction (sinus histiocytos) in lymph nodes in the hepatoduodenal ligament in the plastic stent group. Also, the overall postoperative complication rates were increased in the group where plastic stents were used (72 % vs 52 %), as were the frequency of anastomotic leakages (12 % vs 3.7 %), but none of these postoperative complications reached statistical significance.

Conclusion: Prophylactic antibiotics in ERCP lower the overall complication rates but not sufficiently to recommend this as prophylaxis in every ERCP procedure. On the contrary, prophylactic pancreatic stents could be used more frequently in ERCP and larger diameters and longer stents demonstrated lower complications rates. We also launched an ERCP complexity grading scale (HOUSE) and validated it in relation to complication rates and procedure duration. Finally, we demonstrated that SEMS could be used in resectable periampullary tumors and found no differences in bacterial growth in intraoperatively collected bile but a lower preoperative stent dysfunction rate if SEMS were used. Neither did we find any intraoperative technical downsides when using SEMS, or any disadvantages in postoperative complication rates.