Figure. The omental flap was pulled through between the posterior surface of the pancreaticojejunostomy and portal vein (A). It was rolled over the anterior surface of the pancreaticojejunostomy and anchored by several stitches for fixing and attaching the omentum to the pancreaticojejunostomy site (B).
Choi SB, Lee JS, Kim WB, Song TJ, Suh SO, Choi SY. Efficacy of the Omental Roll-up Technique in Pancreaticojejunostomy as a Strategy to Prevent Pancreatic Fistula After Pancreaticoduodenectomy. Arch Surg. 2012;147(2):145-150. doi:10.1001/archsurg.2011.865
Author Affiliations: Department of Surgery, Korea University Guro Hospital, Korea University College of Medicine, Seoul.
Background Most morbidity and mortality are caused by a pancreatic fistula after pancreaticoduodenectomy (PD), and its prevention is the major concern. We applied the omental roll-up technique around pancreaticojejunostomy and investigated the effectiveness of this technique to prevent a pancreatic fistula.
Design Retrospective study.
Setting Tertiary hepatobiliary and pancreas surgery clinic, Korea University Guro Hospital, Seoul.
Patients Between March 1, 2009, and March 31, 2011, 68 patients underwent PD. The patients were divided into 2 groups according to the surgical application of the omental roll-up technique around the PJ site: group 1 (those who did not undergo the omental roll-up technique) compared with group 2 (those who did undergo the omental roll-up technique).
Main Outcome Measure The occurrence of a pancreatic fistula.
Results No differences were noted in the clinical characteristics, including patients' demographics and operation-related factors, between the 2 groups. A pancreatic fistula occurred in 23 of 39 patients in group 1 (59%) and in 6 of 29 patients in group 2 (20.7%). Group 2 had a significantly lower incidence of pancreatic fistula (P = .002), and these fistulas were classified as being grade A using the International Study Group on Pancreatic Fistula Definition showing a transient high amylase level in the drainage fluid without significantly affecting the patient's recovery. Drain removal was performed earlier in group 2 (P < .001). Mean postoperative hospital stay was 23.4 days in group 1 compared with 15.9 days in group 2 (P = .009). Overall mortality was 1.5%; however, no deaths were related to a pancreatic fistula.
Conclusions The omental roll-up technique for the PJ site definitely reduced the occurrence of a pancreatic fistula. Therefore, the omental roll-up technique is a simple and effective strategy to prevent a pancreatic fistula.
Pancreaticoduodenectomy (PD) is the treatment choice for benign or malignant tumors located in the periampullary region. Due to the complexity of the anastomosis, especially in pancreaticojejunostomy (PJ), there is a possibility of developing life-threatening complications, and reported morbidity and mortality rates are still high.1- 8 The specific complications associated with the PD procedure are delayed gastric emptying, pancreatic fistula, and postpancreatectomy hemorrhage.
Several methods have been introduced to prevent delayed hemorrhage from the pseudoaneurysm caused by pancreatic leakage. Wrapping the stump of the gastroduodenal artery using the falciform ligament,3 or omental flaps for protection and isolation of splanchnic vessels from the PJ site,9 has been reported to decrease the risk of postoperative intra-abdominal bleeding following PD.
The omental roll-up technique in PJ was not performed until 2009 in the Department of Surgery, Korea University Guro Hospital, Korea University College of Medicine, Seoul. However, after experiencing several episodes of major bleeding after PD in consecutive patients, the omental roll-up technique around the PJ site was introduced in May 2010, to isolate the vessels from the PJ site and to prevent pancreatic leakage from the PJ site. Therefore, the objective of this study was to evaluate the efficacy of the application of a simple omental roll-up technique around the PJ site to prevent a pancreatic fistula and postpancreatectomy hemorrhage.
Between March 2009 and March 2011, 68 patients underwent PD for benign or malignant disease in the Department of Surgery, Korea University Guro Hospital, Korea University College of Medicine. The following preoperative demographics and clinical information were retrospectively obtained from the patients' medical records: age, sex, symptoms, comorbidities, preoperative laboratory results, American Society of Anesthesiologists class (http://www.asahq.org/clinical/physicalstatus.htm), operative procedures, operative time, transfusion requirements, postoperative pathologic results, interval from operation date to initiation of enteral feeding and to removal of drainage catheter, postoperative complications, and length of hospital stay.
All 68 PDs were performed by one experienced staff surgeon (S.Y.C.), affiliated to the Department of Surgery at our institute. In most of the patients, pylorus-preserving PD was performed rather than the classic Whipple procedure (2 of 68 patients). In the patients with periampullary malignant neoplasms, lymph node dissection was performed routinely around the common hepatic artery, porta hepatis, celiac axis, and portocaval and retropancreatic areas. Pancreaticojejunostomy was performed with a double-layered, end-to-side, duct-to-mucosa anastomosis between the main pancreatic duct and jejunal wall. The outer layer was composed of the remnant pancreatic parenchyma and the seromuscular layer of the jejunum. The inner-layer anastomosis of the pancreatic duct and jejunal mucosa was performed by continuous suture with 6-0 monofilament polyglyconate (Maxon; Covidien) or 5-0 polydioxanone monofilament. An internal stent was inserted into the pancreatic duct and jejunum for all patients who underwent PJ, except for those who had a dilated pancreatic duct (5 of 68 patients) or those in whom PJ was performed by the Dunking method (1 of 68 patients). The end-to-side hepaticojejunostomy was carried out using continuous 4-0 or 5-0 polydioxanone monofilament sutures.
For omental roll-up around the PJ site, the omental flap was made from the greater omentum that was the most redundant fold and was located close to the PJ site. The omentum was divided longitudinally, along the avascular area preserving 1 or 2 omental branches of gastroepiploic vessels. The omental flap was pulled through between the posterior surface of the PJ and portal vein and was rolled over the anterior surface of the PJ. The rolled omentum was anchored by several silk stitches between the omentum and seromuscular layer of the jejunum and pancreatic parenchyma, respectively, for fixing and attaching the omentum to the PJ site (Figure). Then, the reconstruction was completed using an end-to-side antecolic gastro/duodenojejunostomy. Two surgical drains were routinely placed anterior and posterior to the PJ site and were exteriorized on the right side of the abdominal wall. All the patients received subcutaneous injection of octreotide acetate, 100 μg, 3 times daily during the first postsurgical week. The nasogastric tube was removed on postoperative day 1. If the patient showed good recovery, an enteral diet was permitted following a release of gas. However, if the patient showed signs of pancreatic leakage, initiation of oral feeding was decided on by considering the patient's condition and amount of the drainage fluid.
The patients were divided into 2 groups according to the surgical application of the omental roll-up technique around the PJ site: group 1 (those who did not undergo the omental roll-up technique) and group 2 (those who did undergo the omental roll-up technique). The clinical outcomes, including the postoperative complications, were compared between the 2 groups.
Postoperative pancreatic fistula has been defined according to the International Study Group on Pancreatic Fistula (ISGPF) as a drain output of any measurable volume of fluid on or after postoperative day 3 with an amylase content greater than 3 times the upper normal serum amylase value.10 Pancreatic fistulas were also classified into 3 grades defined according to ISGPF clinical criteria: grades A, B, and C.10 In addition, in patients with low amylase levels of the drainage fluid that could not reflect the clinical findings of pancreatic fistulas, including abdominal pain and distention with impaired bowel function, fever (>38°C), and abruptly elevated serum leukocyte count, the presence of pancreatic fistula was confirmed on abdominal computed tomography in several cases by identifying an abnormal fluid collection or abscess around the PJ site. When an attempt to aspirate an abnormal fluid collection detected on computed tomography resulted in drainage of amylase-poor fluid (which is not associated with pancreatic fistula), it was categorized as an intra-abdominal abscess (drainage of thick and purulent fluid) or fluid collection (drainage of serous or serosanguineous fluid).
Postpancreatectomy hemorrhage represented all the postoperative episodes of hemorrhage and was classified according to the International Study Group of Pancreatic Surgery (ISGPS) definition on the basis of the time of onset, location and cause, and severity.11Delayed gastric emptying has been defined as an inability of the patient to return to a standard diet by the end of the first postoperative week with an interval of prolonged nasogastric tube intubation.12
Data are expressed as mean (SD). Statistical calculations were performed using SPSS (version 13.0 for Windows; SPSS, Inc). Comparisons between groups were tested using the Pearson χ2 test. For continuous variables, independent samples t test was used to compare the 2 groups. P < .05 was considered statistically significant.
Of 68 patients, 44 were men (64.7%) and 24 were women (35.3%) with a mean age of 64 years (age range, 22-82 years). Twenty patients (29.4%) were asymptomatic preoperatively and were diagnosed with periampullary tumor on medical check-up. Most common symptoms were jaundice (n = 25 [36.8%]) and pain (n = 14 [20.6%]). Forty-seven patients (69.1%) had 77 preoperative comorbidities, such as hypertension (n = 38), diabetes mellitus (n = 25), chronic hepatitis B (n = 3), valvular heart disease (n = 2), chronic renal failure (n = 2), stomach cancer (n = 2), congestive heart failure with atrial fibrillation (n = 1), angina pectoris (n = 1), stroke (n = 1), idiopathic thrombocytopenic purpura (n = 1), and polymyositis (n = 1). Forty-two patients (61.8%) underwent preoperative biliary drainage for decompression of biliary obstruction.
Mean operation time was 265 minutes, ranging from 170 to 420 minutes. Combined superior mesenteric vein and portal vein resection was performed in 5 patients due to vascular invasion of the tumor. The omental roll-up technique was applied for protection around the PJ site in 29 patients (42.6%), and PJ without omental roll-up was performed in the remaining 39 patients (57.4%). Thirty-nine patients (57.4%) required intraoperative transfusion. Mean postoperative period until drain removal was 14 days. Mean interval between operation and the initiation of oral feeding was 4 days. There were 18 cases (26.5%) of pancreatic cancer, 19 cases (27.9%) of distal bile duct cancer, 17 cases (25.0%) of cancer of the ampulla of Vater, and 14 cases (20.6%) of benign pancreatic disease. Mean tumor size was 2.7 cm.
Fifty-nine postoperative complications occurred in 37 patients, with an overall postoperative morbidity of 54.4% (Table 1). Postoperative pancreatic fistulas developed in 29 patients (42.6%), and the grades of pancreatic fistulas according to the ISGPF definition10 were as follows; 14 of 29 patients (48.3%) had grade A, 9 patients (31.0%) had grade B, and 6 patients (20.7%) had grade C. None of the patients having pancreatic fistula required surgical intervention; however, 6 patients received percutaneous drainage by an interventional radiologist. Other pancreatic fistulas resolved spontaneously after conservative management.
Five patients (7.4%) developed postpancreatectomy hemorrhage. Early-onset (≤24 hours after the end of the index operation), extraluminal hemorrhage (grade A and B postpancreatectomy hemorrhage according to the ISGPS definition11) occurred in 2 patients, and these patients were treated conservatively. One patient showed late, extraluminal hemorrhage compatible with grade B postpancreatectomy hemorrhage and was controlled by angiographic embolization. Two patients demonstrated late, life-threatening extraluminal hemorrhage and were controlled by angiographic embolization (grade C). Delayed gastric emptying developed in 4 patients (5.9%), and the grade of delayed gastric emptying according to the ISGPS definition12 was as follows: 2 patients were grade A; 1, grade B; and 1, grade C. Complications directly related to the omental roll-up technique around the PJ site did not occur.
One patient developed massive cerebrovascular embolism postoperatively and died at postoperative day 39, and the overall mortality was 1.5%. However, no deaths were related to pancreatic fistula.
No differences were noted in age, sex, American Society of Anesthesiologists class, body mass index, preoperative biliary drainage, and preoperative laboratory results between the 2 groups (Table 2). The 2 groups were not statistically different (Table 3) for operation-related factors, including operative time, insertion of the PJ stent, size of the pancreatic duct, pancreatic texture, and combined vascular resection/need for transfusion.
Group 2 (those who underwent the omental roll-up technique) had a significantly lower incidence of pancreatic fistula (P = .002). Pancreatic fistulas occurred in 23 of 39 patients (59.0%) in group 1 and in 6 of the 29 patients (20.7%) in group 2. In group 2, of the 6 patients who developed pancreatic fistulas, most of them (5 of 6) were classified as grade A without having any significant effect on the patients' recovery. Group 1 had a marginally significant higher incidence of higher-grade pancreatic fistula (grades B and C) than group 2 (P = .054). No statistically significant difference was noted in the occurrence of other complications except pancreatic fistula between the 2 groups. There was no difference in the incidence of postpancreatectomy hemorrhage between the 2 groups. However, 2 patients in group 1 had severe late pseudoaneurysm bleeding that was suspected to be associated with a pancreatic fistula. Drain removal was performed earlier in group 2 than in group 1 (10.5 days vs 16 days, P < .001). Mean postoperative hospital stay was 23.4 days in group 1 compared with 15.9 days in group 2 (P = .009) (Table 3).
Recent advances in surgical techniques and adequate management of postoperative complications have led to improved clinical outcomes of PD, and the mortality following PD has decreased below 5%.2,4,5 However, the morbidity following PD is still high and has been reported to be approaching 50% to 60%.1- 5 As is well known, most of the morbidity and mortality are caused by pancreatic fistula. Septic condition associated with a grade C pancreatic fistula might result in the patient's death.7 Local inflammation caused by pancreatic fistula erodes the walls of major vessels skeletonized after lymph node dissection near the pancreatic bed, and may result in pseudoaneurysm formation, sloughing of a ligature at an arterial stump, or both.8 Therefore, late postpancreatectomy hemorrhage mainly caused by postoperative pancreatic fistula still remains a fatal complication, with a reported mortality of 30% to 50%.6- 8
The incidence of postoperative pancreatic fistula after PD has been reported to vary from 6.7% to 53%.2,3,9,13- 18 In the literature review, the definition of pancreatic fistula between studies is somewhat different, thereby explaining the variance of occurrence of pancreatic fistula. The incidence of pancreatic fistula in the current study seemed to be high. However, the criteria of grade A pancreatic fistula according to the ISGPF was strict with, transient high-amylase level in the drainage fluid. Half the overall pancreatic fistulas were classified into grade A in the current study without clinical impact. Similarly, for the group of patients in whom the omental roll-up technique was applied, most of the pancreatic fistulas were classified as grade A, having little clinical significance. Death attributable to a pancreatic fistula did not occur during the current study.
The significant risk factors, including soft pancreatic parenchyma,7,16,18,19 small pancreatic duct size,13,16 requirement of blood transfusion,7,18 postoperative bleeding,7 presence of coronary artery disease,17,18 and older age,16,18 have been reported to be related to the development of a pancreatic fistula. Besides these patient- and disease-related clinical factors, factors related to operative techniques have also been investigated and reported.13,15,20- 23
Prevention of a pancreatic fistula is the major concern following PD, and pancreatic surgeons have made great efforts to reduce this almost unavoidable complication. Many technical attempts and studies have been performed, and controversies still remain between the studies. At our institute, almost all anastomosis of the remnant pancreas was performed by continuous duct-to-mucosa PJ, using an internal stent. Some authors have reported that continuous duct-to-mucosa anastomosis is superior to interrupted duct-to-mucosa anastomosis in reducing the incidence of pancreatic fistula.20 Pancreaticogastrostomy has been reported to be a contestable alternative to PJ in decreasing the incidence of pancreatic fistula.21,22 In a prospective randomized trial, external drainage of the pancreatic duct with a stent has been reported to reduce the incidence of pancreatic fistula in the stent group compared with the no-stent group.13 Furthermore, both internal and external drainage were safe techniques in PJ; however, internal drainage simplifies postoperative management and shortens the postoperative hospital stay.15 The omental wrap or flap has also been used to prevent the catastrophe associated with a pancreatic fistula by protecting the major vessels.2,3,9,24,25
In an experimental study, omental wrapping of an avascular anastomotic segment in the small intestine of dogs showed a transition of vessels from the omentum to the intestine to preserve the avascular segment.26 Omental wrap has been reported to be protective in an unsafe anastomosis by providing a biological viable plug to prevent leakage and a source of granulation tissue and neovascularization for wound repair.27 On the basis of these experimental studies, omental roll-up around the PJ site can be expected to have the role of a patch to accelerate adhesion and to reinforce and protect the PJ site. In one study, 2 omental flaps each for PJ and duodenojejunostomy were suggested to decrease the risk of postpancreatectomy hemorrhage related to pancreatic fistula; however, the difference in the occurrence of pancreatic fistula and intra-abdominal bleeding between the group using the omental flap and the group not using the omental flap was not statistically significant.2 Maeda et al9 reported that wrapping an omental flap around dissected splanchnic vessels was performed in 100 cases of PD, and it reduced postpancreatectomy hemorrhage but failed to prevent pancreatic fistulas.9 The omental flap was placed cranially through the space between the PJ site and the portal vein covering the splanchnic vessels, to separate the PJ site from the vessels that is the possible source of postpancreatectomy hemorrhage. However, the anterior portion of the PJ site was not covered by the omental flap, which was different from the omental roll-up technique used in our study. In our study, the omental flap was placed between the PJ site and portal vein and rolled up around the PJ site covering its anterior portion, and the omentum was well fixed and attached effectively to the PJ site with several anchoring sutures. Therefore, we could definitely reduce the incidence of pancreatic fistula. Furthermore, in group 2, most of the patients who developed pancreatic fistulas were classified as having grade A without having any significant clinical impact. Other studies have also suggested and reported the effectiveness of the omental wrapping or flap for preventing serious conditions following the development of pancreatic fistulas.24,25
Falciform ligament was also used for a similar purpose of wrapping the stump of the gastroduodenal artery or skeletonized vessels for prevention of delayed postpancreatectomy hemorrhage following PD3 or distal pancreatectomy.14 However, the omentum is more redundant and abundant, and therefore a larger patch can be available for the PJ site than the falciform ligament. The falciform ligament might be adequate to separate the splanchnic vessels from the PJ site, but it is insufficient for rolling over the PJ site, especially in slender patients. No complications were caused in the current study by the omental flap procedure per se, such as portal vein compression or omental necrosis that promotes intra-abdominal infection, which is consistent with the results of other studies.9
In conclusion, the omental roll-up technique for the PJ site definitely reduced the occurrence of pancreatic fistula in the current study. For the group of patients in whom the omental roll-up technique was applied, most of the pancreatic fistulas were classified as grade A, having little clinical significance. It is a simple and easy to apply technique without any risk related to the procedure. Although the number of patients in this study was small, it can be concluded that the omental roll-up technique is a simple and effective strategy to prevent pancreatic fistula. We believe that the omental roll-up around the PJ with several anchoring sutures for attaching the omentum to the PJ site might promote healing of the PJ by adhesion. Omental roll-up can also have an important role in the separation and protection of the splanchnic vessels from the pancreatic fistula to prevent postpancreatectomy hemorrhage.
Correspondence: Sang Yong Choi, MD, PhD, Department of Surgery, Korea University Guro Hospital, Korea University College of Medicine, 80, Guro-dong, Guro-gu, Seoul, Korea (email@example.com).
Accepted for Publication: August 5, 2011.
Author Contributions:Study concept and design: S. B. Choi, Kim, Song, Suh, and S. Y. Choi. Acquisition of data: S. B. Choi, Lee, and S. Y. Choi. Analysis and interpretation of data: S. B. Choi, Lee, and S. Y. Choi. Drafting of the manuscript: S. B. Choi and S. Y. Choi. Critical revision of the manuscript for important intellectual content: S. B. Choi, Kim, Song, Suh, and S. Y. Choi. Statistical analysis: S. B. Choi, Lee, Kim, and S. Y. Choi. Administrative, technical, and material support: S. B. Choi, Kim, and S. Y. Choi. Study supervision: Song, Suh, and S. Y. Choi.
Financial Disclosure: None reported.