Postoperative Complications Linked to Pancreaticoduodenectomy. An Analysis of Pancreatic Stump Management Enrico Benzoni, Aron Zompicchiatti, Enrico Saccomano, Dario Lorenzin, Umberto Baccarani, Gianluigi Adani, Luigi Noce, Alessandro Uzzau, Carla Cedolini, Fabrizio Bresadola, Sergio Intini Department of Surgery, University of Udine, School of Medicine, Udine, Italy. Abstract Introduction Aims. To analyze the role of different procedures in the The treatment of pancreatic cancer is undertaken with two management of pancreatic stump according to the incidence different objectives. The ﬁrst is radical surgery for patients of postoperative morbidity derived from the data of a single with early stage of disease, mainly stage I and partially II. center surgical population. Methods. From 1989 to 2005 In all other cases, the aim of treatment is the palliation of we performed 76 pancreaticoduodenectomies (PD) and 26 the distressing symptoms related to this cancer. distal pancreatectomies (DP). The surgical reconstruction Despite the advances made in surgical technique and after PD was as follows: 11 manual non-absorbable stitches perioperative care, limited progress has been made in closure of the main duct, 24 closures of the main duct with improving the survival of patients with this disease. Five-year linear stapler, 17 occlusions of the main duct with neoprene survival rates are 5% to 20% for patients undergoing potentially glue and 24 duct-to-mucosa anastomosis. Results. In the PD curative resection; postoperative disease recurrence occurs group, the morbidity rate was 60%, caused by: pancreatic commonly [1-3]. Despite this, surgical resection remains the leakage in 48% of patients, hemorrhagic complications only potentially curative therapeutic option in 10% of patients in 10% following surgical procedure and infectious for whom resection is possible [2, 3]. The most recent series complications in 15%. After DP we recorded: leakage in from institutions specializing in the treatment of pancreatic 3.9%, haemoperitoneum in 15.4% and no complications in cancer report mortality rates after surgical procedure less 80.7%. The multivariate analysis showed that the in-hospital than 5% [1, 2]. But morbidity rates remain high: 30% to mortality was linked to the surgical procedure (PD, p=0.003) 60% [1, 2]. The majority of perioperative complications are and to the following complications: pancreatic leakage not life threatening, though they result in increased length (p=0.004), haemoperitoneum (p=0.00045) and infectious of hospital stay and costs, readmissions for care, and delay complications (p=0.0077). Bleeding complications, biliary in adjuvant therapy. Hemorrhagic complication occurs in anastomosis leakage and infectious complications were 5–16% of patients following pancreaticoduodenectomy (PD); consequences of pancreatic leakage (p=0.025, p=0.025 and pancreatic leakage is a major cause of morbidity and mortality p=0.025 respectively). Conclusion. Manual non-absorbable after PD, with incidence varying between 6 and 24% and a stitch closure of the main duct and occlusion of the main duct mortality rate up to 40% [1, 4]. with neoprene glue should be avoided in the reconstructive The variations in methods of pancreatic stump phase. management and the volume of literature available on both main pancreatic duct and pancreaticoenteric anastomosis leak indicate the concern associated with the leakage and Key words the continuing efforts to prevent it. Pancreaticoduodenectomy - postoperative complication In this study we have analyzed the role of different - anastomotic leak - Whipple’s procedure procedures in the management of pancreatic stump after PD, according to the incidence of postoperative morbidity derived from data in a single center surgical population. J Gastrointestin Liver Dis March 2008 Vol.17 No 1, 43-47 Address for correspondence: Dr. Benzoni Enrico University of Udine Patients and methods Schoool of Medicine P.le S. Maria della Misericordia, From 1989 to 2005, 137 consecutive patients, who 33100 Udine, Italy had undergone surgical interventions for pancreatic e mail : firstname.lastname@example.org cancer, were followed up at our department. Preoperative 44 Benzoni et al studies employed abdominal ultrasonography, abdominal 7th days after resection. All patients underwent an X-ray computed tomography (CT) and endoscopic retrograde chest examination on the 3rd postoperative day, and a cholangiopancreatography (ERCP), with brushing for US examination or a CT scan during the ﬁrst week after cytology. Only a few cases had undergone ﬁne needle resection. transduodenal biopsy during ERCP. Serum tumor markers When the CT scan during the 1st week after resection such as carcinoembryonic antigen (CEA) and carbohydrate recorded a ﬂuid collection around the surgical ﬁeld with antigen 19-9 (CA19-9) were measured preoperatively. The no accompanying symptoms, a percutaneous drainage following biological investigation were also performed was performed to assess the amylase content and for preoperatively: hemoglobin, red blood cell count, white blood microbiological analysis. cell count, hematocrit, platelet count, partial thromboplastin Pancreaticoenteric anastomotic leak was deﬁned as: time, prothrombin time, international normalized ratio, drainage of >10 mL of ﬂuid with high amylase content (>3 serum albumin, serum proteins, and liver function tests times serum amylase level) for > 4 days postoperatively; (serum bilirubin, transaminases, γ-glutamil transpeptidase, or percutaneous drainage of amylase-rich intra-abdominal lactic dehydrogenases). collection/abscess; or as intraoperative demonstration An endoscopic stenting was positioned during ERCP of pancreaticoenteric anastomosis disruption at re- whenever required (obstructive jaundice) in order to reduce exploration. Blood transfusion was considered necessary bilirubin serum level before surgery. when hemoglobin level was less than 8 g/dL and important The ﬁnal assessment of resectability was made during bleeding was considered when a fall in haemoglobin >2 g/dL, surgical intervention considering: retroperitoneal extension haemodynamic instability and/or patient requiring >2 units of the tumors, liver metastases not identiﬁed by CT, of blood transfusion were recorded. Hepaticojejunostomy involvement or encasement of superior mesenteric artery, leak was diagnosed when a drainage of >50 mL of bilious superior mesenteric vein or portal vein. An intraoperative ﬂuid after postoperative day 4 was recorded or a leak was pathological examination was performed on the lymph shown by contrast radiology. Infectious complications, nodes of the inferior vena cava and aorta groups and of any diagnosed when the leucocyte count was >12 000/mmc, group around the pancreatic area if they were tangible. A body temperature >38.5°C and blood culture positive, were pathological examination of the pancreatic, duodenal and treated with antibiotics selected according to blood culture bile duct margin was made to assess the radicality of the results and antibiograms. resection. Lymph nodes were considered positive if they Statistical analysis showed evidence of carcinoma, whether by direct extension Data are expressed as mean ± standard deviation and of tumor or by metastases. Margins were considered positive incidence is reported as percentage; χ2 test was used to if there was histological evidence of adenocarcinoma in compare the mean values of the incidence of complications any of the surgical margins of the resected specimen (i.e., resulting between different types of “management of the pancreatic neck, duodenal, or bile duct margins). duct” during PD. DP patients were assumed as a control group We performed 76 PDs, 26 distal pancreatectomies (DP) for comparing the postoperative complications incidence. and 35 total pancreatectomies. We excluded from the study Multivariate analysis was used to assess independent 35 patients who had undergone total pancreatectomy, and variables, which may inﬂuence survival, in-hospital death we recorded the results of 102 patients who had underwent and occurrence of complications. Signiﬁcant difference was PD or DP. deﬁned as p< 0.05. Data processing was performed using We did not perform pylorus-preserving procedures in this SPSS® 13.0 for Windows® Evaluation Version. series of patients. During DP, the parenchymal transection was performed with a linear stapler (either Tyco® GIA™ 60- 80, or Tyco® TA™ 60-90). The surgical reconstruction after Results PD was as follows: 11 manual non-absorbable stitches closure of the main duct, 24 closures of the main duct with linear Mean age was 62.06 ± 9.81 years. Ninety four patients stapler (Tyco® TA™ 60-90), 17 occlusions of the main duct had ductal adenocarcinoma; 6 of these had adenosquamous with neoprene glue and 24 duct-to-mucosa anastomosis. carcinoma, 24 mucinous cystadenocarcinoma, 2 serous We always placed drains at the time of surgery: one was cystadenocarcinoma. Eight patients had neuroendocrine placed near the suture of the transected parenchyma, if a tumors. Pathologic staging was as follows: Ia - 3.6%, Ib - transection with a linear stapler or a main duct closure was 16%, IIa - 13.9%, IIb - 48.9%, III - 6.6%, IV 11%. In-hospital performed, or near the anastomosis, if a duct-to-mucosa mortality rate was 4.37%. In the PD group, the morbidity anastomosis was performed. Another one was placed near rate was 60%, caused by: pancreatic leakage in 48% patients, the hepaticojejunostomy. haemorrhagic complication in 10% of patients following Postoperative care surgical procedure, and infectious complications in 15%. Starting with the ﬁrst postoperative day, we administered Blood transfusion was requested during 50 PD procedures somatostatin or its analogs in all patients in order to inhibit and the mean blood amount transfused was 1103.43 ± 978.7 pancreatic secretion. cc. No differences among the PD subtype groups in the Blood tests were performed on the 1st, 3rd, 5th and estimated blood loss were recorded. Postoperative complications linked to pancreaticoduodenectomy. 45 According to the type of surgical reconstruction after temporary occlusion of the main duct with neoprene glue PD we recorded the following complications: after manual and 29.1% after duct-to-mucosa anastomosis. non-absorbable stitches closure of the main duct: 27.3% no A signiﬁcant statistical difference, regarding re-operation complications, 27.3% leakage, 18.1% haemoperitoneum, rates, was found between closure of the main duct with 27.3% infections (intra-abdominal abscess concomitant linear stapler vs. temporary occlusion of the main duct with to leakage); after closure of the main duct with linear neoprene glue (p= 0.049) and closure of the main duct with stapler: 50% no complications, 33.3% leakage, 12.5% linear stapler vs. duct-to-mucosa anastomosis (p=0.003). haemoperitoneum, 4.2% infections (wound infections); After DP we recorded: 80.7% no complications (a 40% after occlusion of the main duct with neoprene glue: 32.9% rate was recorded after PD, p=n.s.), 3.9% leakage (a 48% rate no complications, 55.3% leakage, 5.9% haemoperitoneum, was recorded after PD, p=0.03), 15.4% haemoperitoneum 5.9% infections (one case of cholangitis resolved with (a 10% rate was recorded in PD, p=n.s.). antibiotics); after duct-to-mucosa anastomosis: 58.2% no At the multivariate analysis, in-hospital mortality was complications, 29.2% leakage, 4.2% haemoperitoneum, 8.4% linked to the surgical procedure (PD, p=0.003) and the infections (intra-abdominal abscess concomitant to leakage). arising of complications: pancreatic leakage (p=0.004), The statistical differences and incidence of complications haemoperitoneum (p=0.00045) and infectious complication resulting between different procedures of the management of (p=0.0077). Overall complications were linked to: age >70 yr pancreatic stump after PD are shown in Table 1 and Fig. 1. (p=0.0139), T stage 3 (p=0.031) and N stage 2 (p=0.000001), surgical procedure (PD p=0.0018) and pancreatic residual treatment (neoprene glue closure p=0.003 and stapler closure p=0.002). Bleeding complications, biliary anastomosis leakage and infectious complications were consequences of pancreatic leakage (p=0.025, p=0.025 and p=0.025, respectively). Discussion Complications that are related to the pancreatic remnant still represent a substantial risk for death after pancreatic head resections [5-7]. Failure of a surgical anastomosis has serious consequences Fig.1 Incidence of complications. resulting after different particularly in cases of anastomosis of the pancreas to the managements of pancreatic stump after PD; absolute values small bowel, because of the digestive properties of activated (number of patients) upon the columns. See Table 1 for statistical pancreatic secretions. An important factor in the prevention signiﬁcance of pancreatic ﬁstula in patients with a pancreaticojejunostomy is technical precision and gentleness in construction of the Re-operation rates were: 27.2% after manual non- pancreatic anastomosis. Cattell  in 1943 advocated the absorbable stitches closure of the main duct, 23.8% after necessity of a pancreaticoenteric anastomosis because of the closure of the main duct with linear stapler, 6% after high mortality rate when leaving the pancreatic remnant in Table 1. Incidence of complications due to different managements of pancreatic stump after PD (χ2 test) (n.s.: not signiﬁcant) Type of management of pancreatic stump Overall complications Leakage Haemoperitoneum Infectious disease p-value p-value p-value p-value Manual stiches closure of the main duct (11 pts.) vs. <0.05 n.s. n.s. n.s. Neoprene glue closure (17 pts.) Manual stiches closure of the main duct (11 pts.) vs. <0.005 <0.05 n.s. n.s. Main duct closure with stapler (24 pts.) Manual stiches closure of the main duct (11 pts.) vs. <0.001 <0.001 n.s. <0.05 Duct to mucosa anastomosis (24 pts.) Neoprene glue closure (17 pts.) vs. Main duct closure <0.05 <0.05 <0.05 <0.01 with stapler (24 pts.) Neoprene glue closure (17 pts.) vs. Duct to mucosa <0.005 <0.001 <0.05 n.s. anastomosis (24 pts.) Main duct closure with stapler (24 pts.) vs. Duct to <0.05 <0.05 <0.05 <0.01 mucosa anastomosis (24 pts.) 46 Benzoni et al situ, either with or without ligature of the pancreatic duct. Based on our results, bleeding complication, biliary In the history of PD, some authors [9-11] reported that the anastomosis leakage and infectious complications were the absence of an anastomosis to the pancreatic remnant might consequences of pancreatic leakage. prevent a large proportion of postoperative complications. Intra-abdominal abscesses are seen in about 3-10% A pancreatic ﬁstula from the oversewn pancreatic remnant after PD, and a percutaneous drainage is recommended for is less dangerous than one from the pancreaticojejunal collections larger than 5 cm. Antibiotics should be given anastomosis because there is no defect in the small bowel intravenously [23, 24]. and no activation of pancreatic enzymes. Pancreatic ﬁstula is still an unsolved problem, as there Another technique investigated has been obliteration is still no consensus on a uniform deﬁnition of ﬁstula. The closure of the pancreatic duct with a chemical substance, broad range drainage rates reported in the literature (2-28%) thus avoiding a pancreaticojejunostomy. This method was is largely a function of the deﬁnition used, and there is no proposed by Gebhardt et al , who studied the effect of evidence that a management of pancreatic stump after PD occlusion of the pancreatic duct system with Ethibloc, an is better than others [25, 26]. alcohol prolamine, in animal experiments. The pancreatic We recorded a 60% rate of complications in the PD group, duct may also be occluded with a ﬁbrin glue solution, while only 19.3% of patients in the DP group experienced Tissucol®, which was found to have a more protective effect postoperative complications, and in particular only 3.9% of on beta cell function than the other solutions used . DP vs. 48% of PD patients (p=0.03) recorded a pancreatic Di Carlo et al  described lower morbidity and ﬁstula. We believe that a higher duct pressure and volume mortality in a non-randomized trial in 50 patients using duct of pancreatic ﬂuid into the main duct after pancreatic occlusion with neoprene glue after Whipple’s procedure, resection could explain the difference between PD and DP even if a side effect of permanent occlusion, that induces in developing complications, and the same factors could pancreatic atrophy and complete loss of exocrine function, explain the different rates of complications among different was reported. Gail et al  found similar results with approaches to the reconstructive phase. Furthermore, the duct occlusion after Whipple’s operation in patients with blockage of the main pancreatic duct, as recorded in the chronic pancreatitis. Lorenz et al  noted fewer early occlusion with ﬁbrin glue, even transiently, might artiﬁcially complications in oncologic pancreatic surgery, but not in increase the secretion of pancreatic juice in the severed patients with chronic pancreatitis. As recently reported by secondary canals or on the suture line and, consequently, Suc et al , who performed a prospective randomized both yield and support a pancreatic ﬁstula . trial on ﬁbrin glue occlusion on the pancreatic stump, this As claimed by the literature, risk factors for pancreatic technique did not signiﬁcantly decrease the rate of intra- leakage comprise general factors (age, gender, jaundice, abdominal complications, notably pancreatic ﬁstula, or their malnutrition), disease-related factors (pancreatic pathology, severity after pancreatic resection. pancreatic texture, pancreatic duct size, pancreatic juice Recently, Bilimoria et al  demonstrated that the output) and procedure-related factors (operative time, incidence of leakage is reduced signiﬁcantly when the resection type, anastomotic technique, intraoperative blood pancreatic duct is identiﬁed and directly ligated: pancreatic loss). leak rates were 9.6 % in subgroups having duct ligation and Clearly, the high morbidity and mortality rates associated 34.0 % in subgroup having no duct ligation (p<0.001). At with this operation underscore the magnitude and difﬁculty multivariate analysis, failure to ligate the pancreatic duct of this type of procedure. This high leak rate in the different was the only feature associated with an increased risk for PD groups is consistent with reports of other authors and pancreatic leak (p = 0.001). highlights the difﬁculties in managing the pancreatic stump. In the past, a lower morbidity was recorded in the It is interesting to note that the PD groups involving closure/ application of mechanical stapler during the parenchymal occlusion of the pancreatic duct had similar or higher leaks transection , but actually there is accordance in the and overall complication rates when compared to the duct- usage of the stapler and its beneﬁts, and a careful drain to-mucosa anastomosis group. This is an important ﬁnding, management is recommended to achieve a good outcome because patients in whom a duct occlusion is performed in patients with ﬁstula . develop severe pancreatic exocrine insufﬁciency that causes The incidence of bleeding complications after PD malabsorption. ranges from 5 to 16%, and can occur in up to 60% in cases Furthermore, as regards total pancreatectomy, which is of pancreatic leakage [21-24]. The mortality of a bleeding recommended instead of pancreaticojejunostomy in order complication continues to be high, ranging from 30 to 58% not to develop postoperative complications such as leakage, [23, 24]. Jaundice, hepaticojejuno- and pancreaticojejuno it has to be taken into consideration that this procedure is anastomotic leaks are associated with an increased incidence associated with severe endocrine and exocrine abnormalities. of bleeding. The consistency of pancreas, type of pancreatic Therefore a beneﬁt could be achieved only in patients reconstruction, tumor site and size and operative duration affected by insulin dependent diabetes before surgery. have no inﬂuence on the incidence or type of postoperative In conclusion, based on our experience, we make the bleed. Mortality in bleeders was 34% as compared with 3% following recommendations: manual non-absorbable stitch in non-bleeders [23, 24]. closure of the main duct and occlusion of the main duct with Postoperative complications linked to pancreaticoduodenectomy. 47 neoprene glue should be avoided in the reconstructive phase, 240. and a duct-to-mucosa pancreaticojejunoanastomosis should be 12. Gebhardt C, Stolte M, Schwille PO, Zirngibl H, Engelhardt W. preferred. We recommend pancreaticoenteric reconstruction Experimental studies on pancreatic duct occlusion with prolamine. Horm as a standard procedure after pancreaticoduodenectomy, Metab Res Suppl 1983: 9–11. because all pancreatic duct occlusion techniques are 13. Idezuki Y, Goetz FC, Lillehei RC. Late effect of pancreatic duct associated with high rates of pancreatic ﬁstula complicated ligation on beta cell function. Am J Surg 1969; 117: 33–39. with endocrine insufﬁciency. 14. Di Carlo V, Chiesa R, Pontiroli AE, et al . Pancreato-duodenectomy with occlusion of the residual stump by Neoprene injection. World J Surg 1989; 13: 105–110. Conﬂicts of interest 15. Gall FP, Zirngibl H, Gebhardt C, Schneider MU. Duodenal pancreatectomy with occlusion of the pancreatic duct. None to declare. Hepatogastroenterology 1990; 37: 290–294. 16. Waclawiczek HW, Boeckl O, Lorenz D. Pancreatic duct occlusion with ﬁbrin (glue) to protect the pancreatico-digestive anastomosis References after resection of the head of the pancreas in oncologic surgery. 1. Kuhlmann KF, de Castro SM, Wesseling JG, et al. Surgical treatment Langenbecks Arch Chir Suppl Kongressbd 1996; 113: 252–254. of pancreatic adenocarcinoma; actual survival and prognostic factors 17. Suc B, Msika S, Fingerhut A, et al . Temporary ﬁbrin glue occlusion in 343 patients. Eur J Cancer 2004; 40: 549–558. of the main pancreatic duct in the prevention of intra-abdominal 2. Stojadinovic A, Brooks A, Hoos A, Jaques DP, Colon KC, Brennan MF. complications after pancreatic resection prospective randomized trial. An evidence-based approach to the surgical management of resectable Ann Surg 2003; 237: 57–65. pancreatic adenocarcinoma. J Am Coll Surg 2003; 196: 954–964. 18. Bilimoria MM, Cormier JN, Mun Y, Lee JE, Evans DB, Pisters PW. 3. Ahmad NA, Lewis JD, Ginsberg GG, et al . Long term survival Pancreatic leak after left pancreatectomy is reduced following main after pancreatic resection for pancreatic adenocarcinoma. Am J pancreatic duct ligation. Br J Surg 2003; 90: 190-196. Gastroenterol 2001; 96: 2609–2615. 19. Ahren B, Tranberg KG, Andrei-Sandberg A, Bengmark S. Subtotal 4. Yang YM, Tian XD, Zhuang Y, Wang WM, Wan YL, Huang YT. Risk pancreatectomy for cancer: closure of the pancreatic remnant with factors of pancreatic leakage after pancreaticoduodenectomy. World staplers. HPB Surg 1990; 2: 29-35. J Gastroenterol 2005; 11: 2456-2461. 20. Balzano G, Zerbi A, Cristallo M, Di Carlo V. The unsolved problem 5. Kakita A, Yoshida M, Takahashi T. History of pancreaticojejunostomy of ﬁstula after left pancreatectomy: the beneﬁt of cautious drain in pancreaticoduodenectomy: development of more reliable management. J Gastrointest Surg 2005; 9: 837-842. anastomosis technique. J Hepatobiliary Pancreat Surg 2001; 8: 230- 21. Shankar S, Russell RC. Haemorrhage in pancreatic disease. Br J Surg 237. 1989; 76: 863–866. 6. Sakorafas GH, Friess H, Balsiger BM, Buchler MW, Sarr MG. 22. Rumstadt B, Schwab M, Korth P, Samman M, Trede M. Hemorrhage Problems of reconstruction during pancreatoduodenectomy. Dig Surg after pancreaticoduodenectomy. Ann Surg 1998; 227: 236–241. 2001; 18: 363-369. 23. Niedergethmann M, Farag Soliman M, Post S. Postoperative 7. Bassi C, Falconi M, Molinari E, et al . Duct-to-mucosa versus complications of pancreatic cancer surgery. Minerva Chir 2004; 59: end-to-side pancreaticojejunostomy reconstruction after 175-183. pancreaticoduodenectomy: results of a prospective randomized trial. 24. Balachandran P, Sikora SS, Raghavendra Rao RV, Kumar A, Saxena R, Surgery 2003; 134: 766-771. Kapoor VK. Haemorrhagic complication of pancreaticoduodenectomy 8. Cattell RB. Resection of the pancreas, discussion of special problems. ANZ J Surg 2004; 74: 945–950. Surg Clin North Am 1943; 23: 753-766. 25. Bassi C, Butturini G, Molinari E, et al . Pancreatic ﬁstula rate after 9. Shankar S, Theis B, Russell RC. Management of the stump of pancreatic resection. The importance of deﬁnitions. Dig Surg 2004; the pancreas after distal pancreatic resection. Br J Surg 1990; 77: 21: 54-59. 541–544. 26. Kazanjian KK, Hines OJ, Eibl G, Reber HA. Management of 10. Marcus SG, Cohen H, Ranson JH. Optimal management of the pancreatic ﬁstulas after pancreaticoduodenectomy: results in 437 pancreatic remnant after pancreaticoduodenectomy. Ann Surg 1995; consecutive patients. Arch Surg 2005; 140: 849-854. 221: 635–645. 27. Ohwada S, Ogawa T, Tanahashi Y, et al . Fibrin glue sandwich prevents 11. Papachristou DN, Fortner JG. Pancreatic ﬁstula complicating pancreatic ﬁstula following distal pancreatectomy. World J Surg 1998; pancreatectomy for malignant disease. Br J Surg 1981; 68: 238– 22: 494–498.
Pages to are hidden for
"Postoperative Complications Linked to Pancreaticoduodenectomy. An"Please download to view full document