- (2015) Volume 16, Issue 5
Yoshitaka Nakamura1, Akihiro Yamaguchi1, Hiroomi Matsumura1, Azumi Suzuki2, Hiroki Taniguchi1
Departments of 1Surgery and 2Gastroenterology, Kyoto Second Red Cross Hospital, Japan
Received June 22nd, 2015-Accepted July 25th, 2015
Context Transanastomotic internal pancreatic duct stent placement during reconstruction following pancreaticoduodenectomy is thought to be one of the strategies for preventing postoperative pancreatic fistula and widely performed. However complications related to internal stent migration have very rarely been reported. Here we report a rare case of internal pancreatic stent retrograde migration into the left biliary tree through afferent limb of a Roux-en-Y reconstruction after pancreaticogastrostomy following pancreaticoduodenectomy. Case report A seventy-four-year-old woman underwent subtotal stomach preserving pancreaticoduodenectomy with right hemicolectomy for intraductal papillary mucinous neoplasm of the pancreas head and adenocarcinoma of the cecum. Pancreaticogastrostomy was undertaken during subtotal stomach preserving pancreaticoduodenectomy with internal pancreatic stent placement. Post operative course was uneventful and she was discharged on the fifteenth postoperative day. She remained almost asymptomatic after discharge except for slight vague abdominal pain. Twenty-seven days after the operation, an internal pancreatic stent was detected at right upper quadrant in routine abdominal X-ray at the time of outpatient visit. An abdominal computed tomography scan revealed that a stent was migrated into the left hepatic duct trough hepaticojejunostomy. Enteroscopic retrieval of the stent was then successfully undertaken. Conclusions In the present case, an internal pancreatic stent unexpectedly migrated into the hepatic duct against the peristaltic direction of the afferent limb. Although complications related to surgically placed internal pancreatic stent migration have rarely been reported so far, surgeons should be aware of the possibility of internal stent migration and any postoperative imaging studies should be considered focusing on the position of the devices placed intraoperatively.
complications; Pancreaticoduodenectomy; Pancreatic Ducts
CT computed tomography; IPMN Intraductal papillary mucinous neoplasm; PD Pancreaticoduodenectomy; PF Pancreatic fistula; PG Pancreaticogastrostomy; PJ Pancreaticojejunostomy; POPF Postoperative pancreatic fistula; SSPPD Subtotal stomach preserving pancreaticoduodenectomy
Pancreaticoduodenectomy (PD) is the standard surgical procedure for patients with malignant and benign diseases of the pancreatic head and periampullary region. Although advances in surgical techniques and perioperative management have been achieved so far resulting reduced mortality rates after PD to less than 5% in high-volume centers [1,2], the morbidity rates after PD are remain as high as 30 to 50 % [2-4]. The most common complication after PD is pancreatic leaks, in particular, pancreatic fistula (PF). PF from pancreatic anastomosis remains a major source of morbidity, with a reported incidence of around 2 to 30 % and it also significantly contributes to sepsis, abscess, intra-abdominal bleeding, prolonged hospitalization and even mortality [5-9].
Attempting to prevent the development of postoperative pancreatic fistula (POPF) following PD, various surgical reconstruction techniques and management strategies have been proposed for the anastomosis of remaining pancreas to the gastrointestinal tract [7,10-13].
Pancreatic duct stent placement is one of these tactics commonly performed [14] for bridging the anastomosis of the pancreatic remnant to the gastrointestinal tract and may be useful for decompression of the remnant pancreas, and maintaining patency of the main pancreatic duct, thus prevent failure of anastomosis site and leakage of proteolytic enzyme from the pancreas [15]. A pancreatic internal stent is usually presumed to pass spontaneously through the rectum. Complications directly related to surgically placed internal stent in the pancreatic duct have only rarely been reported so far [16-25]. In those reported cases, anterograde migration of a stent was occurred along with the peristaltic direction. We report herein, for the first time as far as we know, an extremely unusual case of internal pancreatic stent migration into the biliary tree retrogradely through the afferent loop against the peristaltic direction.
A seventy-four-year old woman was referred to our hospital for investigation of positive faecal occult blood test and adenocarcinoma of the cecum was detected by colonoscopy. A contrast-enhanced abdominal computed tomography (CT) scan for further examination incidentally revealed a 3.5 cm of intraductal papillary mucinous neoplasm (IPMN) of the pancreas head, as well as known tumor of the cecum.
Subtotal stomach preserving pancreaticoduodenectomy with right hemicolectomy was undertaken. In the operation, pancreaticogastrostomy (PG) was performed with placement of a 5-Fr polyvinyl chloride tube (MD- 41515 pancreatic duct tube; Sumitomo Bakelite Co. Ltd., Tokyo, Japan) across the anastomosis according to the diameter of the pancreatic duct. An internal stent was then anchored to the pancreatic stump parenchyma by a 5-0 monofilament polypropylene suture (Surgipro™ II; Covidien, Mansfield, MA, USA). For insertion of pancreatic stump into the gastric lumen, a corresponding opening was made on the posterior wall of the stomach followed by interrupted sutures between full thickness of the gastric wall and the pancreatic parenchyma using 4-0 monofilament polyglyconate synthetic absorbable sutures (Maxon™; Covidien, Mansfield, MA, USA). After PG was completed, hepaticojejunostomy was made several centimetres distal to the cut end of the jejunum using an end-to side anastomosis by interrupted 4-0 monofilament polyglyconate sutures. The jejunum was transected at 50 cm distal to the hepaticojejunostomy followed by jejunojejunostomy between proximal cut end of the jejunum and 25 cm downstream of distal cut end of the jejunum using a circular stapler (Curved Intraluminal Stapler 25 mm; Ethicon Endo-Surgery, Cincinnati, OH, USA) with side closure. The distal cut end of the jejunum was closed with a linear stapler (Linear Cutter 75 mm; Ethicon Endo-Surgery, Cincinnati, OH, USA) and then sideto- side gastrojejunostomy was completed with the linear stapler. A 9-Fr polyurethane feeding catheter (Kngaroo™ Jejunostomy catheter; Covidien, Mansfield, MA, USA) was inserted from afferent limb and top of the tube was placed distally for postoperative enteral nutrition. An 8mm silicone drainage tube (Pleats drain tube; Sumitomo Bakelite Co. Ltd., Tokyo, Japan) was placed as shown in Figure 1.
Figure 1. (a.). Reconstruction after subtotal stomach preserving
pancreaticoduodenectomy (SSPPD), (b.). or SSPPD with right
hemicolectomy.
Red arrow shows internal pancreatic stent placed across the
pancreaticogastrostomy.
Yellow arrows show possible route of the stent migration in the present
case.
Postoperative course was uneventful and she was discharged on the 15th postoperative day. She remained almost asymptomatic after discharge except for slight vague abdominal pain. Twenty-seven days after surgery, an internal pancreatic stent was detected at the right upper quadrant in routine abdominal X-ray at the time of outpatient visit (Figure 2). She was not presenting with fever or abdominal pain. On physical examination, she did not appear jaundice and her abdomen was soft and flat, without tenderness. The rest of her examination was unremarkable. Laboratory examination revealed a white blood cell count of 7,800/mm3 (normal range, 3,200 to 9,200), haemoglobin of 7.8 g/dL (normal range, 11.3 to 15.3), platelet cell count of 371,000/mm3 (normal range, 139,000 to 363,000), C-reactive protein 1.53 mg/ dL (normal range, 0.00 to 0.40), amylase 69 U/L (normal range, 40 to 140), total bilirubin 0.3 mg/dL (normal range 0.2 to 1.0), alkaline phosphatise 477 (normal range, 125 to 350), AST 24 U/L (normal range, 10-35) and ALT 23 U/L (normal range, 5 to 40). An abdominal CT scan revealed that a stent was migrated into the left biliary tree through hepaticojejunal anastomosis (Figure 3). Dilatation of the intrahepatic bile duct was not evident.
Figure 2. An abdominal X-ray on the 27th days after surgery. White arrow shows an internal pancreatic stent.
Figure 3. A contrast-enhanced abdominal computed tomography scan. Red arrows show the migrated pancreatic stent in coronal images.
To retrieve the migrated pancreatic stent, enteroscopy was performed. The migrated pancreatic stent was seen protruding into the left biliary tree trough the hepaticojejunostomy. The stent was successfully retrieved with grasping forceps without any complications (Figure 4). After the removal of migrated stent, her vague abdominal pain was resolved and she has been well for 9 months without any evidence of recurrent disease.
Figure 4. Endoscopic retrograde cholangiography during enteroscopic retrieval of the migrated stent. White arrows show the migrated pancreatic stent.
Advances in surgical techniques and perioperative management have made it possible to reduce the mortality rates to less than 5 % after PD, however complication rates remains as high as 30 to 50 %. Most of the significant complications are related to POPF from pancreaticodigestive anastomosis. POPF after PD have been reported in 8 to 29 % of patients and can lead to sepsis, abscess, intra-abdominal bleeding, prolong hospitalization, and even death [6-9].
Attempting to reduce the incidence of pancreatic leakage, various surgical techniques were proposed, including end-to-side pancreaticojejunostomy (PJ) [10], duct-to mucosa anastomosis [7], binding PJ and the use of external stents [11]. Pancreaticogastrostomy (PG) has also been advocated in the past few years, as recent randomized clinical trials suggested the lower postoperative mortality and PF rates than PJ [12,13]. Moreover, PG is thought to be superior to PJ in terms of observation of the remnant pancreas [26]. For those reasons, performing PG during reconstruction following PD is a usual procedure at our institution.
Pancreatic stent placement across the pancreatic anastomosis following PD may be useful for the diversion of pancreatic juice from the pancreatic anastomotic site, decompression of the remaining pancreas, and maintaining patency of the main pancreatic duct [15]. Roder et al. reported a prospective study showing that pancreatic fistula decreased from 29 % to 7 % using an external pancreatic drainage tube after PD [6,22]. Some randomized controlled trials have focused on the use of a stent in the pancreatic duct after PD [27,28] and they concluded that the use of an external stent through the pancreatic anastomosis reduced the PF rate. However, external stents are uncomfortable for the patients and have the potential risk for unexpected removal and there may be problems including twisting, bleeding, and occlusion of the stent [8,29]. External stents are usually removed several weeks after surgery [6]. Despite lack of evidence supporting the advantage of an internal stent to an external stent in terms of the rate of PF, internal stent placement appears to simplify postoperative management and even shorten the postoperative hospitalization after PD and have been receiving much attention. Tani et al. reported that the median postoperative hospital stay in their group with internal drainage was shorter than that in the group with external drainage, although there was no difference in the incidence of PF [15].
An internal pancreatic stent is generally thought to passes spontaneously through the rectum. Yoshimi et al. reported the timing of the defecation of an internal pancreatic stent placed across the PJ, describing that all internal short stents placed in 11 patients who underwent PD followed by end-to-side PJ had fallen out spontaneously by the 176th postoperative day [30]. However, internal pancreatic stents could migrate inward or outward from the position they initially placed. A previous report suggested that proximal migration occurs in 5 to 6 % of the patient, which can lead to obstruction, pancreatitis, or ductal damage, and while distal migration occurs in 8 to 12 % of the cases [31]. Kadowaki et al. reported that the median interval between stent placement and stent defecation and cumulative defecation rate at 1 year were 454 days and 41 %, respectively, in their series of 57 patients who had performed reconstruction with the modified Child method following PD or with the Traverso method following pylorus-preserving PD [22]. They also found the pancreatic stent migration in 7 patients, including 4 cases of migration into the hepatic duct and 3 cases of migration into afferent limb. Although there have been no reported cases of internal stent migration after PG during PD so far as we know, only a few reported complications related to pancreatic stents migration after PJ were found in the literature including proximal migration causing pancreatitis and steatorrhea [16,18,22,23], distal migration into the small intestine presenting as liver abscess [19,20,22], acute cholangitis [22], bezoar ileus [17], vague abdominal pain [21] and even resulting in bowel perforation [24,25]. In those reported cases, the migration site was downstream of the pancreatic anastomosis. It is reasonable that a pancreatic stent detached from the pancreatic anastomosis is carried toward downstream of digestive tract along with the peristaltic direction. In the cases of migration into the biliary tree, the hepaticojejunostomy was distal to the pancreatic anastomosis. Kadowaki et al. suggested that reconstruction in which the hepaticojejunostomy is not distal to the PJ, including Whipple’s method, might be an effective method to avoid migration of the stent into the bile duct following PD [22]. In the present case, hepaticojejunostomy is far from the pancreatic anastomosis and not directly distal to the pancreatic anastomosis. During the period between 2009 and 2014, a total of 132 patients underwent SSPPD in our department, including 2 cases of SSPPD with right hemicolectomy, all of whom underwent the same type of reconstruction as shown in Figures 5a, b respectively. However, in our limited experience, there have been no complications related to internal pancreatic stent migration so far, except for the present case. Although causes of the pancreatic stent migration remain to be obscure, several possibilities could be proposed. Bowel obstruction might be one of the causes for stent migration. If the obstruction occurred distal to the jejunojejunostomy, detached stent could not pass through the obstructed digestive tract and might remain in the jejunum. This could contribute to the increasing opportunity for the stent migration into the afferent limb. However, in our case, ileus or constipation was not evident in the postoperative course. Another possibility is that the feeding jejunostomy catheter might disturb normal peristaltic movement as a foreign body, resulting in the stent migration into the afferent limb. Or, another possibility is the stent that had already been detached from the pancreatic anastomosis might migrate into the afferent limb along with the jejunostomy catheter during the removal of the catheter. In the present case, the pancreatic stent was detected at the lower part of the abdomen in X-ray examination on the 13th postoperative day, suggesting early detachment of pancreatic stent. The jejunostomy catheter was then removed on the next day. This fact may support the hypothesis.
Figure 5. Enteroscopic retrieval of the stent. (a.). A pancreatic stent was intruding to the left hepatic duct through the hepaticojejunostomy. (b.). Migrated stent was retrieved with grasping forceps. (c.). Migrated pancreatic stent after removal.
Despite a lack of evidence regarding the true rate of complications related to surgically placed internal pancreatic stent migration, surgeons should be aware of potential complications induced by stent migration. Moreover, any postoperative imaging studies should be considered focusing on the position of the devices placed intraoperatively. Additionally, X-ray fluoroscopy during the removal of surgically placed devices should be considered.
The authors declare that they have no conflicts of interest.
