eTable 1. Parameters for Patients With POPF Grade B
eTable 2. Postoperative Complications in Patients Who Underwent Distal Pancreatectomy
eTable 3. Risk Factors for Clinically Relevant Pancreatic Fistula
Customize your JAMA Network experience by selecting one or more topics from the list below.
Jang J, Shin YC, Han Y, et al. Effect of Polyglycolic Acid Mesh for Prevention of Pancreatic Fistula Following Distal Pancreatectomy: A Randomized Clinical Trial. JAMA Surg. 2017;152(2):150–155. doi:10.1001/jamasurg.2016.3644
Is polyglycolic acid mesh effective in preventing postoperative pancreatic fistula after distal pancreatectomy?
In this randomized clinical trial of 97 patients (44 in the polyglycolic acidgroup and 53 in the control group), the rate of clinically relevant postoperative pancreatic fistula (grade B or C) was significantly lower in the polyglycolic acid group than in the control group (11.4% vs 28.3%).
Wrapping of the cut surface of the pancreas with polyglycolic acid mesh can significantly reduce the rate of clinically relevant postoperative pancreatic fistula.
The rate of postoperative pancreatic fistula (POPF) after distal pancreatectomy ranges from 13% to 64%. To prevent POPF, polyglycolic acid (PGA) mesh was introduced, but its effect has been evaluated only in small numbers of patients and retrospective studies.
To evaluate the efficacy of PGA mesh in preventing POPF after distal pancreatectomy.
Design, Setting, and Participants
Prospective randomized clinical, single-blind (participant), parallel-group trial at 5 centers between November 2011 and April 2014. The pancreatic parenchyma was divided using a stapling device; no patient was given prophylactic octreotide. Perioperative and clinical outcomes were compared including POPF, which was graded according to the criteria of the International Study Group For Pancreatic Fistulas. A total of 97 patients aged 20 to 85 years with curable benign, premalignant, or malignant disease of the pancreatic body or tail were enrolled (44 in the PGA group and 53 in the control group).
Patients in the PGA group underwent transection of the pancreas and application of fibrin glue followed by wrapping the PGA mesh around the remnant pancreatic stump.
Main Outcomes and Measures
The primary end point of this study was the development of a clinically relevant POPF (grade B or C by the International Study Group grading system). The secondary end point was the evaluation of risk factors for POPF.
The study therefore evaluated a total of 97 patients, 44 in the PGA group and 53 in the control group. Thirty-nine patients were women and 58 patients were men. There were no differences in mean (SD) age (59.9 [12.0] years vs 54.5 [14.1] years, P = .05), male to female ratio (1.0:1.3 vs 1.0:1.7, P = .59), malignancy (40.9% vs 32.1%, P = .37), mean (SD) pancreatic duct diameter (1.92 [0.75] mm vs 1.94 [0.95] mm, P = .47), soft pancreatic texture (90.9% vs 83.0%, P = .17), and mean (SD) thickness of the transection margin (16.9 [5.4] mm vs 16.4 [4.9] mm, P = .63) between the PGA and control groups. The rate of clinically relevant POPF (grade B or C) was significantly lower in the PGA group than in the control group (11.4% vs 28.3%, P = .04).
Conclusions and Relevance
Wrapping of the cut surface of the pancreas with PGA mesh is associated with a significantly reduced rate of clinically relevant POPF.
clinicaltrials.gov Identifier: NCT01550406.
Patients undergo distal pancreatectomy for various benign and malignant conditions of the distal pancreas including chronic pancreatitis (24%), benign pancreatic cystadenoma (22%), and pancreatic adenocarcinoma (18%).1 Since the Efficacy of Stapler Versus Hand-sewn Closure After Distal Pancreatectomy (DISPACT) trial2 reported that there was no significant difference in formation of postoperative pancreatic fistula (POPF) after distal pancreatectomy between stapler closure and hand-sewn anastomosis of the pancreatic remnant, stapler devices have become more common for distal pancreatectomy. In addition, the rate of laparoscopic surgery for distal pancreatic disease has increased, resulting in reductions in postoperative pain and hospital stay.
Postoperative pancreatic fistula remains the leading cause of morbidity after distal pancreatectomy, with a frequency of 13% to 64%.1,3,4 The International Study Group for Pancreatic Fistulas (ISGPF) has defined POPF as a drain fluid amylase concentration greater than 3 times the upper normal serum concentration after postoperative day (POD) 3.5 Methods used to prevent POPF include main pancreatic duct ligation,6 stapled closure,2,7 pancreatico-intestinal anastomosis,8 biologic glues,9,10 reinforcement,11,12 and medications like somatostatin analogue.13-18 However, to our knowledge, none has shown advantages over the others.
Mesh reinforcement is a method of wrapping the remnant pancreatic stump after distal pancreatectomy. Among the meshes introduced is polyglycolic acid (PGA) (Neoveil; Gunze) mesh, a bioabsorbable recombinant membrane made of a synthetic polymer with a cellulose-like structure. Retrospective studies have shown that the rates of POPF with and without PGA mesh reinforcement were 5.6% to 27% and 38.9% to 42%, respectively.11,12 However, the protocols of those studies did not involve standardized operations and did not control for variables associated with POPF. Therefore, this prospective randomized clinical trial was designed to evaluate the ability of PGA mesh reinforcement to prevent POPF after distal pancreatectomy.
This was a randomized, clinical, single-blind (participant), parallel-group trial. The study complied with the Declaration of Helsinki and was approved and overseen by the institutional review board of each participating hospital (Seoul National University Hospital, Gangnam Severance Hospital, Seoul National University Bundang Hospital, Severance Hospital, and Samsung Medical Center). This study was registered at clinicaltrials.gov (NCT01550406). The formal trial protocols can be found in Supplement 1.
Patients were included if they (1) were 20 to 85 years of age; (2) had an life expectancy of at least 24 months; (3) had potentially curable benign, premalignant, or malignant disease of the pancreatic body or tail, as shown by preoperative imaging (computed tomography, magnetic resonance imaging, and/or positron emission tomography); and (4) provided written informed consent. Patients were excluded if they (1) had a chronic pancreatitis-induced atrophic pancreas or calcification of pancreatic parenchyme; (2) were receiving immunosuppressive therapy; (3) had previously undergone chemoradiotherapy; or (4) were known to be substance abusers.
This trial was performed by experienced and specialized pancreatic surgeons from 5 tertiary referral hospitals, each of which performed more than 100 pancreatectomies per year and submitted case report forms without rule violation.
After confirming patient eligibility, patients were randomized 1:1 to undergo stapled transection of the pancreas followed by wrapping the remnant pancreatic stump with PGA mesh (PGA group) or to undergo stapled transection alone (control group). Randomization was performed using an Internet-based system just before surgery. The allocation sequence was computer-generated and randomly stratified by the surgeon.
After randomization, all clinical and pathologic data, including operation field photographs, were stored in a central database (http://en.bri.snuh.org/pub/_/singlecont/view.do). All serious adverse events were submitted to the Clinical Trials Unit, Seoul National University Hospital, Seoul, Korea, and evaluated periodically by an independent data and safety monitoring board blinded to patient treatment.
The consensus on surgical procedures was determined during a meeting of the participating surgeons before the study was started. Patients in the PGA group underwent pancreatectomy, and PGA mesh was tightly wrapped around the pancreatic stump where fibrin glue was first applied followed by stitching with peripancreatic soft tissue (Figure 1). The type of cartridge was based on the thickness and texture of the pancreas and was selected by each surgeon during the operation. Suture reinforcement of the transection margin was not performed. Before closure, 1 closed suction drain was inserted into the bed of the removed portion of the pancreas and maintained for at least 3 days postoperatively to prevent intraabdominal fluid collection and identify POPF. Perioperatively, none of the patients were allowed use of prophylactic somatostatin analogues to prevent POPF development. Each patient was allowed sips of water on POD 1 and a soft blended diet on POD 2. Postoperative assessment included repeated measurements of amylase concentrations in serum and drainage fluid while the drain was in place. All patients underwent abdominal computed tomography (CT) on POD 5 and 3 months after surgery. Drains were removed in patients without abnormal CT findings on PODs 5 to 7. Removal of the drain was delayed in patients with postoperative fluid collection and/or in need of additional drainage. Postoperative pancreatic fistula was defined as a drain fluid amylase concentration greater than 3 times the upper normal serum concentration after POD 3 as defined by ISGPF criteria.5 If CT scan showed drain migration, the grade of POPF was determined using other parameters, including abdominal pain, distention with impaired bowel function, fever (temperature >38°C), serum leukocyte count greater than 10 000 cells/mm3, increased C-reactive protein concentration, and peripancreatic fluid collection, rather than from amylase concentration in drainage fluid.5
The primary end point of this study was the development of a clinically relevant POPF (grade B or C by the ISGPF grading system). The study sample was calculated by assuming that the clinically relevant POPF rates would be less than 3.8% and 27% in the PGA and control groups, respectively, based on retrospective findings.12 Enrolling 106 patients would therefore provide 80% power to detect such a difference with an adjusted significance level α = .025, β = 0.2, and an expected dropout rate of 23%, as determined by the Medical Research Collaborating Center, Seoul National University Hospital, using Power Analysis and Sample Size software (NCSS LLC). The secondary end point was the evaluation of risk factors for POPF.
Results were presented as mean (SEM). Patient demographics and clinical characteristics were compared using the χ2 test or Fisher exact test for categorical variables and t test and the Mann-Whitney test for continuous variables. In assessing risk factors associated with POPF, only variables statistically significant by univariate analysis were included in the multivariate analysis, which was performed using logistic regression. All statistical analyses were performed using SPSS, version 21.0 (SPSS Inc), with P values less than .05 considered statistically significant.
Between November 2011 and April 2014, a total of 107 patients were enrolled, with 49 randomized to the PGA group and 58 randomized to the control group. After enrollment, 10 patients were excluded because of (1) use of somatostatin analogue (n = 4), (2) surgical rule violation (n = 4), (3) withdrawal of consent after initial agreement (n = 1), or (4) loss to follow-up (n = 1). The study therefore evaluated a total of 97 patients, 44 in the PGA group and 53 in the control group (Figure 2).
Pathologic indications for distal pancreatectomy are listed in Table 1. Of the 97 patients, 28 (28.9%) underwent surgery for pancreatic ductal adenocarcinoma (PDAC). The incidence of POPF varied according to different pathologic diagnosis. The rate of POPF was lowest in patients diagnosed as having PDAC (35.7%, P = .002).
There were no significant differences between the PGA and control groups in mean age, sex distribution, histology, pancreatic main duct diameter, soft pancreatic texture, thickness of pancreatic transection margin, operation time, estimated blood loss, and postoperative hospital stay (Table 2). Most patients in each group underwent simultaneous splenectomy, including 35 of 44 (79.5%) in the PGA and 42 of 53 (79.2%) in the control group. Fifty-eight patients (59.8%) were diagnosed as having any grade POPF and 20 patients (20.6%) with clinically relevant POPF (ISGPF grade B or C). The grading of POPF is listed in Table 3. The rates of any grade POPF did not differ significantly in the PGA and control groups (29 of 44 patients [65.9% ] vs 29 of 53 patients [54.7% ]; P = .26). Although the rate of ISGPF grade A was significantly higher in the PGA group than in the control group (26.4% vs 54.5%; P = .02), the rate of clinically relevant POPF (grade B or C) was significantly higher in the control group than in the PGA group (28.3% vs 11.4%, P = .04). Parameters for patients with POPF grade B decided according to ISGPF criteria are shown in eTable 1 in Supplement 2.
eTable 2 in Supplement 2 includes a detailed list of other complications. Complications were observed in 26 patients (59.1%) in the PGA group and 30 patients (56.6%) in the control group. There was no evidence of increased complications secondary to POPF except for fluid collection. There was no operation-related mortality during observation period.
Risk factors for clinically relevant POPF were analyzed (eTable 3 in Supplement 2). Univariate analysis showed that younger age and being in the control group were associated with clinically relevant POPF. Multivariate analysis showed that age younger than 60 years was the only statistically significant risk factor for clinically relevant POPF (odds ratio [OR], 7.23; 95% CI, 1.93-27.12; P = .003). Lack of PGA wrapping after transection of the pancreatic parenchyme was not significant (OR, 2.85; 95% CI, 0.89-9.11; P = .08).
Methods of preventing POPF following distal pancreatectomy have included stapled closure, use of fibrin sealant on the pancreatic stump, mesh reinforcement, and medications such as somatostatin analogue. Although a randomized clinical trial18 showed that perioperative treatment with pasireotide significantly reduced the rate of clinically significant POPF, leak, or abscess (9% vs 21%, P = .006), pasireotide is very costly.18 Because PGA mesh is strong, pliable, and easily inserted in large abdominal wall defects, it was previously used successfully to bridge abdominal wall defects and prevent evisceration or herniation after laparotomy. Polyglycolic acid mesh induces an inflammatory reaction immediately after insertion and is infiltrated by granulation tissue within 3 weeks. Two to 3 months after insertion, the material is absorbed. Several retrospective studies have evaluated the ability of PGA mesh reinforcement to prevent POPF. For example, pancreatic leak was observed in 10 of 38 patients (26%) undergoing conventional resection with suture ligation of the pancreatic duct or nonreinforced stapled resection, compared with 2 of 47 patients (4%) undergoing stapled resection using PGA mesh reinforcement.11 Another study12 found that application of PGA felt to stapled pancreatic transection lines was significantly associated with reduction in the occurrence of POPF (3.5% vs 22%, P = .04). Moreover, use of PGA felt with a fibrin sealant was found to significantly reduce the risk of severe POPF (4% vs 27%, P = 0.02).12
This multicenter, prospective randomized clinical trial showed that wrapping of the pancreas cut surface with PGA mesh was associated with significant reduction in the rate of clinically relevant POPF (P = .04). Another randomized clinical trial19 at a single center found that PGA mesh reinforcement of the pancreatic transection line was associated with significant reduction in the incidence of significant POPF in patients undergoing distal pancreatectomy (P = .007).19
Although wrapping the pancreas cut surface with PGA mesh after distal pancreatectomy cannot prevent biochemical pancreatic leakage (grade A POPF), it reduced the rate of clinically relevant POPF without any additional morbidities. Following intraperitoneal injection into mice, degraded PGA mesh was reported to induce acute peritonitis, characterized by neutrophil infiltration.20 Clinically relevant POPF in this study may thus have been prevented by chronic inflammation and adhesion around adjacent tissues associated with the acute peritonitis induced by PGA mesh.21,22
Our analysis of risk factors for POPF showed that the rate of clinically relevant POPF was significantly lower in patients older than 60 years than those younger than 60 years, similar to previous findings.23,24 Watanabe et al23 reported that the regeneration of pancreatic acini after pancreatic resection was shown to be markedly attenuated in the pancreas of older individuals through mice tests, most likely because of decreased PI3K/Akt activation.23 In addition, older patients were more likely to have PDAC found within a fibrotic gland in this study (mean [SD] age, 63.3 [9.3] years vs 54.4 [14.0] years, P = .002). The rate of POPF was relatively lower in patients with PDAC or intraductal papillary mucinous neoplasm. Pancreatic ductal adenocarcinoma or intraductal papillary mucinous neoplasm tended to occur in older patients compared with other tumors. This may explain the lower rate of POPF after distal pancreatectomy in older patients.
Of the 28 patients (28.9%) diagnosed as having PDAC, only 13 patients (13.4%) had a firm or hard pancreatic texture. Of the 13 patients, 10 patients were diagnosed as having PDAC. Pancreatic texture depends on the location of the main mass. For example, if the main mass is located in the distal body or tail, the texture of the head or neck can be soft despite the final diagnosis as PDAC. Many patients with PDAC (18 of 28, 64.3%) had soft pancreatic texture because pancreases were resected at the normal texture proximally to the main mass.
None of the patients in this study were treated with a somatostatin analogue postoperatively. Although a 2014 trial showed that perioperative pasireotide significantly reduced the risk and severity of POPF among patients undergoing pancreatic resection,18 somatostatin analogues were excluded from the protocol of this study to avoid confounding factors.
The study has some limitations. First, type of the cartridge to be used during transection of the pancreas was determined by each surgeon’s preference. Although there were no reports about significant differences in rate of POPF according to types of the cartridge, it is possible for them to affect the rate of POPF after transection of the pancreas. It may be a reason why thickness of pancreatic transection is not a significant risk factor in this study. Thus, another trial to solve this problem is ongoing with respect to the relationship between type of cartridge and POPF. Second, the rate of POPF after distal pancreatectomy was higher than in previous studies. This may have been caused by our inserting 1 closed suction drain into all patients to identify postoperative complications and routinely measure amylase concentrations on POD 1, 3, and 5. In addition, although radiologic documentation is neither mandatory nor necessarily recommended for diagnosis of POPF,5,25 we performed routine postoperative abdominal CT scans of all patients to determine postoperative complications. It is considered that grade A POPF may be difficult to be detected without detailed analyses of prospective data collected after distal pancreatectomy including postoperative drain amylase concentrations and abdominal CT scans such as in this study.
Wrapping of the cut surface of the pancreas with PGA mesh was associated with a significantly reduced rate of clinically relevant POPF. Application of PGA mesh may become a routine procedure to prevent POPF following distal pancreatectomy.
Corresponding Author: Sun-Whe Kim, MD, PhD, Department of Surgery, Seoul National University College of Medicine, 101 Daehak-ro, Chongno-gu, Seoul 110-744, South Korea (firstname.lastname@example.org).
Accepted for Publication: July 2, 2016.
Published Online: October 26, 2016. doi:10.1001/jamasurg.2016.3644
Author Contributions: Dr Jang had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Jang and Shin contributed equally to this study.
Concept and design: Jang, Park, H. Hwang, Yoon, D. Hwang, C. Kang, Heo, S. Kim.
Acquisition, analysis, or interpretation of data: Jang, Shin, Y. Han, Park, H. Han, H. Hwang, D. Yoon, J. Kim, Y. Yoon, Lee, M. Kang, Y. Chang, J. Chang, Jung.
Drafting of the manuscript: Jang, Shin, Y. Han, D. Hwang, Heo, J. Chang.
Critical revision of the manuscript for important intellectual content: Jang, Shin, Park, H. Han, H. Hwang, D. Yoon, J. Kim, Y. Yoon, C. Kang, Lee, Y. Kang, Y. Chang, J. Chang, Jung, S. Kim.
Statistical analysis: Jang, Shin, Y. Han, D. Hwang, J. Chang.
Administrative, technical, or material support: Jang, Y. Han, Park, H. Hwang, D. Yoon, J. Kim, D. Hwang, C. Kang, M. Kang, Y. Chang, J. Chang, Jung, S. Kim.
Study supervision: Jang, Park, H. Han, H. Hwang, D. Yoon, Lee, Heo, M. Kang, S. Kim.
Conflict of Interest Disclosures: None reported.