Prevention and treatment of pancreatic fistula after pancreatic body and tail resection: current status and future directions

Postoperative pancreatic fistula (POPF) is the most common and critical complication after pancreatic body and tail resection. How to effectively reduce the occurrence of pancreatic fistula and conduct timely treatment thereafter is an urgent clinical issue to be solved. Recent research standardized the definition of pancreatic fistula and stressed the correlation between POPF classification and patient prognosis. According to the literature, identification of the risk factors for pancreatic fistula contributed to lowering the rate of the complication. Appropriate management of the pancreatic stump and perioperative treatment are of great significance to reduce the rate of POPF in clinical practice. After the occurrence of POPF, the treatment of choice should be determined according to the classification of the pancreatic fistula. However, despite the progress and promising treatment approaches, POPF remains to be a clinical issue that warrants further studies in the future.


Introduction
Distal pancreatectomy (DP) is a standard surgical procedure for removing pancreatic body and tail lesions. In recent years, the prognosis of benign and malignant diseases of the body and tail of the pancreas has significantly improved by the advancement of surgical methods and proposal of new concepts. Nevertheless, the incidence of postoperative complications (especially pancreatic fistula) remains high, which has been troubling clinical surgeons all along. Therefore, how to prevent the occurrence of postoperative pancreatic fistula (POPF) and standardize the treatment of choice have attracted research interest in recent years.
This review aims to provide an overview of the recent advances in the field of prophylaxis and management of pancreatic fistula after laparoscopic DP and discuss potential future developments.

Definition of pancreatic fistula
Pancreatic fistula can be diagnosed on the basis of clinical manifestations and the level of amylase in drainage fluid. However, it should be differentiated from other diseases with similar clinical manifestations, such as intestinal fistula, bilioma, abdominal abscess, and intestinal obstruction [1,2]. In light of the confusion over the definition and classification of pancreatic fistula, the International Study Group of Pancreatic Fistula (ISGPF) issued the diagnostic criteria for pancreatic fistula in 2005 [3]; that is, the peritoneal drainage fluid after operation can be measured, in which the amylase concentration exceeds the upper limit of normal serum amylase level by three times or more than 300 IU/L. These criteria have been widely accepted. The ISGPF further updated the definition of pancreatic fistula in 2016 [4], emphasizing the correlation between pancreatic fistula and clinical prognosis in addition to the original definition.
The ISGPF classifies pancreatic fistula into two categories (biochemical pancreatic fistula and clinically relevant pancreatic fistula) and three grades (A, B, C) according to the clinical manifestations, imaging manifestations, drainage duration, necessity of reoperation, pancreatic fistula-related death, infection signs, sepsis, and re-admission. Class A pancreatic fistula is now referred to as biochemical pancreatic fistula because there are no clinical symptoms. Class B and C pancreatic fistulas are defined more strictly than before. Class B pancreatic fistulas require significant clinical interventions such as continuing drainage for more than three weeks or endoscopic and percutaneous drainage [5]. Class C pancreatic fistulas refer to pancreatic fistulas requiring further surgical intervention. Patients with class C pancreatic fistulas have organ failure, which can lead to death.

Risk factors and prediction of pancreatic fistula
Identifying the risk factors of pancreatic fistulas and actively predicting and preventing them in clinical practice are important. Literature reports indicate that patient characteristics, operation approaches, and techniques are correlated to the occurrence of pancreatic fistula (Table 1).
Previous studies that reported the risk factors of pancreatic fistula focused almost exclusively on the operation-related index, such as long operation time (more than 480 min), intraoperative blood loss of more than 1000 mL, and combined splenectomy [6][7][8]. All of them were listed as the risk factors of pancreatic fistula after operation. With the development of surgical techniques and experience, researchers have gradually excluded the factors of surgical techniques and sought to find possible risk factors from patient characteristics; for example, BMI > 25 is generally considered as an independent risk factor for pancreatic fistula after operation [9][10][11].
Recent studies tend to focus on pancreatic features per se. An analysis of 41 patients by Arai et al. [12] showed that pancreatic thickness ( > 15 mm) was an independent risk factor for pancreatic fistula after pancreatectomy. Consistently, a retrospective analysis by Chang et al. in 2017 [13] predicted the risk of pancreatic fistula by measuring the thickness and sectional area of pancreas with imaging technology. They concluded that thicker pancreas and larger pancreatic cross-section are risk factors for pancreatic fistulas after operation. The reason may be that a thick pancreas is predisposed to insufficient stapled disclosure during clamping, and large residual pancreatic volume leads to the preservation of most pancreatic exocrine function. In the same year, Fukuda et al. [14] determined the thickness of the pancreas, pancreatic index (CT value of pancreas/CT value of spleen), and the ratio of pancreatic thickness to pancreatic index by using digital imaging technology. They verified that pancreatic thickness was an independent risk factor for pancreatic fistulas after DP through multivariate analysis. Most recently, an analysis of 157 patients by Sivesri et al. [15] showed that the pancreatic division level did not affect the rate of POPF after comparing the outcomes of DP by body/tail division (body-tail group) with those by neck division (neck group). Concerning the tissue texture of the pancreas and the diameter of the pancreatic duct, soft tissues and small pancreatic ducts are regarded as culprits for pancreatic fistulas after pancreaticoduodenectomy; however, no significant relationship between them and pancreatic fistula after DP was observed [16]. The results also implied that the mechanism of pancreatic fistula after the two procedures was different.
With the rapid development of imaging technology in recent years, computed tomography and magnetic resonance imaging have become more effective in the diagnosis and classification of various pancreatic diseases, thereby provide more guidance for the surgical methods of choice and help to reduce the occurrence of pancreatic fistulas and other complications [17,18].

Prevention and management of pancreatic fistulas
Since the pancreatic fistula after operation often lead to catastrophic consequences, how to find appropriate ways to reduce the occurrence of POPF has become a hot topic in recent years ( Table 2).

Management of pancreatic stump
Closure of the pancreatic stump directly -transition from handsewn to stapled closure Closing the pancreatic stump properly is the most critical step to prevent the occurrence of pancreatic fistula. Manual suture is a classical approach for the closure of pancreatic stump; however, the incidence of pancreatic fistula remains high from traditional handsewn technique. In recent years, with the rapid development of minimally invasive technology, surgical instruments, biological materials, and other surgical techniques have been constantly improved and optimized.

Management of the main pancreatic duct
Unlike pancreaticoduodenectomy, pancreatic body-tail resection retains the function of the Oddi sphincter, which leads to high intrapancreatic duct pressure. This explains why the incidence of pancreatic fistula after DP is significantly higher than that after pancreaticoduodenectomy [19]. Bilimoria et al. [20] retrospectively analyzed 126 cases of pancreatic body and tail resection in 2003. They found that the incidence of pancreatic fistula in the pancreatic duct ligation group was significantly lower than that in the non-ligation group (9.6% vs. 34%). Multivariate analysis also indicated that non-ligation of the pancreatic duct was an independent risk factor for POPF, with some studies supporting this view [7,11]. By contrast, the series from Ferrone et al. [21] in 2008 failed to confirm the prophylactic role of transfixion of the pancreatic duct for fistula development. Thus, a definitive conclusion cannot be reached given these inconsistent findings.
Due to the rapid implementation of minimally invasive DP, stapled closure of the pancreas has emerged as the preferred method for parenchymal division, which leads to the difficulty of finding and ligating the pancreatic duct during operation. Nevertheless, while laparoscopic DP is performed in many centers, the operation of open DP in the same period mostly includes finding the main pancreatic duct and ligating it [22][23][24].
Application of pancreatic stump-covering material During DP, the hepatic round ligament and falciform ligament can be freed and sutured through the lesser omentum to reinforce the closure of pancreatic stump. The buttressing of pancreatic closures with autoplasty has been reported mostly as anecdotal evidence [24][25][26][27]. A randomized controlled trial (RCT) by Carter et al. [26] in 2012 randomized 109 patients into two groups: closure of the pancreatic remnant with or without falciform ligament patch and fibrin glue. The trial failed to show any benefit of falciform ligament patch and fibrin glue in terms of pancreatic fistula in DP. To address the inconsistence in the literature, Hassenpflug et al. [27] completed a relevant randomized controlled clinical study (the DISCOVER Randomized Controlled Trial) in 2016. In their study, the research group was treated only by hepatic round ligament and falciform ligament covering and fixing the closure of the pancreatic stump, without additional bioglue treatment. The incidence of grades B and C pancreatic fistulas in the study group was 10% lower than that in the control group although no statistical difference was observed. The multivariate regression analysis showed that the use of this autologous material to cover the pancreatic stump was one of the protective factors to prevent the occurrence of clinically relevant pancreatic fistula. Chen et al. [28] also believe that this method can help to reduce the occurrence of pancreatic fistula, and further clinical studies may be needed in the future to address this issue in depth and standardize the operation process.
In addition to the hepatic round ligament, other autologous tissues with serous surface have been proposed to cover the pancreatic stump. In Oláh et al.'s study [29], 70 patients were randomly divided into two groups: the pancreatic stump was closed by staples alone in one group (35 cases), whereas part of the jejunal loop was freed and covered the stapled pancreatic stump with jejunal serosa suture in the other group (35 cases). The total POPF rate in the jejunal-covering group was low, but it did not affect the clinical process (no difference was observed between the two groups in terms of grades B and C fistulas). In 2015, Fujii et al. [30] proposed an improved Blumgart stump closure method: the pancreatic stump was embedded in the seromuscular layer of the jejunum after the elevated jejunum (EJ) was free. The results showed that this method could reduce the incidence of POPF, and multivariate analysis proved that this EJ embedding approach was an independent protective factor for POPF. In addition, Wang et al. [31] proposed to cover the pancreatic stump with the greater omentum during DP, which can reduce the incidence of pancreatic fistula and is less expensive, practical, and safe.
At present, the following biomaterials are used for closing the pancreatic stump: fibrin glue with or without material covering, absorbable gasket suture wrapping pancreatic stump, synthetic gasket assisted reinforced staplers. Recent randomized controlled clinical studies have confirmed that the use of absorbable fibrin glue (TachoSil) to cover (without suture) the pancreatic stump cannot reduce the incidence of pancreatic fistula [23,[32][33][34][35]. Matsumoto et al. reported in 2016 [36] that the use of absorbable Vicryl gasket suture to cover the pancreatic stump can help reduce the risk of pancreatic fistula after DP. An RCT exploring the efficacy of polyglycolic acid (PGA) mesh for the prevention of pancreatic fistula following DP was conducted by Jiang et al. [37] in the same year. In this study, patients in the PGA group underwent transection of the pancreas. After application of the fibrin glue, the PGA mesh was wrapped around the remnant pancreatic stump. The results of this trial offered high-quality evidence supporting that wrapping of the cut surface of the pancreas with PGA mesh was associated with a significantly reduced rate of clinically relevant POPF [37].
Although the effect of biological materials on the prevention of pancreatic fistula is still controversial, further research on improving suture technology combined with more reasonable and effective new materials for the prevention of pancreatic fistula is still promising.

Stapled closure of pancreatic remnant
Given the wide application of laparoscopic techniques in pancreatic surgery and the continuous advancement of surgical instruments in recent years, laparoscopic and robot-assisted approaches are increasingly being used for DP. However, whether stapled closure can replace traditional manual suture still needs to be elucidated. In addition, the safety and impact of stapled approach on postoperative complications are unknown. Researchers once found that stapled closure of pancreatic stump resulted in a higher incidence of pancreatic fistula compared to classical manual suture [6,38]. This phenomenon may be related to the imperfect design of early staplers and inexperience of surgeons. Diener et al. [39] conducted a double-blind randomized controlled clinical trial of 352 patients in 21 European hospitals (DISPACT) in 2011. In this trial, 352 patients were divided into the stapled closure group (177 cases) and manual suture group (175 cases). The results showed no significant difference in the incidence of pancreatic fistula and mortality between the two groups. In the following years, reports indicated that stapler transaction provided equal rate of pancreatic fistula compared with handsewn approach, or even lower rate of clinically significant fistula. This finding was further supported by a meta-analysis in 2015, which included 191 patients with stapled closure and 190 patients with handsewn closure. The fistula rates were 34% for the stapled group and 41% for the handsewn group, indicating no statistical difference (P = 0.66) [40].
Slow parenchymal flattening technique was first described in 2008 by Okano et al. [41].In this technique, the closure jaw was clamped carefully and slowly, taking more than 5 min at a fixed speed, with the jaws of the stapler being held shut for 2 min after firing [41]. Nakamura et al. reported a similar prolonged peri-firing compression technique in 2011 [42]. Both techniques were formulated to reduce the pancreatic fistula by avoidance of tearing and destruction of the pancreatic capsule and parenchyma during transaction. This finding was further supported by the study of Hirashita et al. [43] in 2018. This approach has been widely accepted in recent studies comparing the handsewn and stapled closure of the pancreatic remnant, which yielded no difference between the two approaches in terms of POPF [10,22].
In recent years, most authors advocated the use of synthetic gasket-assisted reinforced stapler to strengthen the closure of pancreatic stump to prevent POPF occurrence. An RCT in 2012 by Hamilton et al. [44] found that mesh reinforcement of stapled pancreatic transection decreases the incidence of pancreatic leakage for DP. In the series, the authors used mesh reinforcement of the staple line with either SEAMGUARD ® or Peri-Strips Dry ® . ISGPF grades B and C fistulas were found in 1.9% (1/53) of patients undergoing resection with mesh reinforcement and 20% (11/45) of patients without mesh reinforcement (P = 0.0007). Most recently, however, another RCT by Kondo et al. [45] failed to show any advantages of mesh reinforcement in terms of POPF over stapled closure without mesh. In the study, they used PGA (Neoveil) as the mesh, a bioabsorbable recombinant membrane made of a synthetic polymer. No significant difference in the incidence of clinically relevant pancreatic fistula was observed between the reinforced stapler and stapler groups (16.3% vs. 27.1%, P = 0.15). Further highpower studies with various kinds of meshes could help to clarify this topic in the future.
The stapler size of choice remained to be an issue during laparoscopic distal pancreatectomy (LDP). Kim et al. found that there is no suitable-sized cartridge for a pancreas thicker than 12 mm [46]. Most recently, Dokmak et al. [47] reviewed 130 patients who underwent stapler closure during left pancreatectomy. Multivariate analysis suggested that the stapler cartridge size did not influence the development of clinically significant POPF.

Pancreatic-digestive tract anastomosis
The anastomosis of pancreatic stump with the intestine or stomach can avoid hemorrhage and necrosis of abdominal tissue and organs caused by pancreatic fistula from pancreatic stump to the abdominal cavity. Moreover, it can retain the drainage of pancreatic fluid to the digestive tract to maintain pancreatic exocrine function. Pancreaticdigestive tract anastomosis was first used for digestive tract reconstruction after pancreaticoduodenectomy. In 2011, Sudo et al. [48] first proposed the anastomosis of the pancreatic stump with the posterior gastric wall during DP. Among the 21 patients who underwent this operation, no grades B and C pancreatic fistulas were observed, and the incidence of grade A pancreatic fistulas was 21%. Yanagimoto et al. [49] also reported reduced POPF by pancreaticogastrostomy during DP compared with closure of the pancreatic stump. However, inconsistent results have been reported in literature. Klein et al. [50] retrospectively analyzed 198 cases of pancreatic body and tail resection in 2012. They found no significant difference in the incidence of pancreatic fistula between the pancreatic stumpdigestive tract anastomosis group and the stump closure group. A randomized controlled clinical study by Kawai et al. [51] in 2016 showed that pancreaticojejunostomy did not reduce the incidence of pancreatic fistula after DP compared with the stump closure approach. A multi-center randomized controlled clinical study by Uemura et al. [52] in 2017 compared 36 cases of pancreaticogastrostomy with 37 cases of manual suture of pancreatic stump after DP. The results showed that the operation time of pancreaticogastrostomy group was longer, but the incidence of peritoneal effusion was lower. No significant difference in the incidence of pancreatic fistula and other complications were observed between the two groups. To address this inconsistence within the literature, pancreaticojejunostomy was applied in 18 patients in our institute to manage the pancreatic remnant after DP during the past two years. In this series, we introduced a new technique: invaginated end-to-end pancreaticojejunostomy with transpancreatic transverse U-sutures [53]. Of note, only two patients developed grade A-type POPF, and no patients developed grade B/C-type POPF. The result of this pilot study seems promising. Overall, whether the management of pancreatic stump-digestive tract anastomosis can reduce the incidence of pancreatic fistula may depend on the detail technique of the approach.

Other intraoperative management
To reduce the incidence of pancreatic fistula, some researchers have proposed to use special instruments, such as radiofrequency energy instruments or LigaSure, to treat pancreatic stump besides conventional surgical scalpel, electrotome, ultrasound scalpel, and stapler to cut the pancreas. However, these instruments are still limited to animal experiments or sporadic small-sample clinical reports [54][55][56][57].
A prospective multi-center study in 2015 [58] revealed that the routine use of intraperitoneal drainage during DP increased the rate of pancreatic fistula formation. Another multi-center randomized controlled study reported by Van Buren et al. [59] in 2017 showed that routine placement of abdominal drainage after pancreatectomy had no effect on the frequency of postoperative imaging, percutaneous drain placement, reoperation, readmission, or quality of life scores of patients after DP, but was related to the occurrence of peritoneal effusion. More recently, Seykora et al. found that early drain removal after DP is associated with better outcomes compared with late removal and no drain placement for POPF [60].

Regulation of intrapancreatic pressure during perioperative period
In light of the correlation between pancreatic duct pressure and the occurrence of pancreatic fistula, some researchers proposed to take measures to reduce intrapancreatic pressure to prevent POPF. In a non-randomized prospective cohort study from 2010, Rieder et al. [61] found that no pancreatic fistula occurred in 25 patients who underwent preoperative pancreatic duct stent placement, whereas five (22%) of 23 patients who did not receive stent placement developed pancreatic fistula after DP. This finding indicated that preoperative pancreatic duct stenting can effectively prevent POPF. In a later series, Hashimoto et al. [62] retrospectively analyzed 223 cases of pancreatic body and tail resection and found that preoperative endoscopic pancreatic duct stenting is a protective factor that could reduce the incidence of pancreatic fistula. By contrast, a randomized controlled clinical study by Frozanpor et al. [63] in 2012 analyzed the use of pancreatic duct stents before pancreatic body and tail resection and showed that preventive drainage by pancreatic duct stents did not reduce the incidence of pancreatic fistula after pancreatic surgery. The authors believe that this may be due to the inflammation of pancreatic duct caused by stent placement, which leads to increased pancreatic duct pressure. To avoid the impact of stent-related complications on pancreatic fistula, Hackert et al. [64] completed a prospective clinical trial in 2017. They injected a smooth muscle relaxant into the pancreatic duct before operation. Preliminary results confirmed that this procedure could reduce the incidence of POPF after DP.

Other clinical treatments, such as drug use and diet control
Octreotide has been widely used in pancreatic diseases and post-pancreatic operation. Although the overall incidence of pancreatic fistula can be reduced, no evidence suggests that octreotide can effectively prevent the occurrence of clinically relevant pancreatic fistula [65]. In recent years, a new somatostatin analog, pasireotide, has been reported to effectively reduce the incidence of pancreatic fistula after DP [66,67].
Opioid analgesics have long been used in postoperative analgesia, especially after pancreatic surgery. Recent reports recommended to reduce the use of narcotics as opioid analgesics can cause Oddi sphincter contraction, which will increase intrapancreatic pressure and lead to pancreatic fistula after surgery [11,64].
Nowadays, preoperative neoadjuvant therapy (chemotherapy and/or radiotherapy) has been increasingly used in patients with pancreatic cancer. Literature reports [68,69] show that preoperative neoadjuvant therapy can reduce the incidence of pancreatic fistula after operation due to the change of pancreatic tissue texture and the inhibition of pancreatic exocrine function. Antibiotic treatment may benefit the outcome of DP; positive result of drainage culture was identified as a new independent risk factor for POPF, and detection of microorganisms in the fluid collection of POPF resulted in high morbidity and mortality [70,71].
Comprehensive management should be taken for the prevention and treatment of pancreatic fistula after operation. The physical condition of preoperative patients, risk factor assessment, and management after operation have a great impact on the incidence of POPF. At the same time, once pancreatic fistula occurs, timely diagnosis and effective treatment are the key factors affecting patient prognosis.

Comprehensive management of pancreatic fistula of DP
The treatment of choice for POPF is determined by the PF grades [7]. Most pancreatic fistulas belonging to A and B grades could be managed through conservative treatments, such as nutritional support, somatostatin and its analogs, and sufficient drainage [6,28,29]. However, some grades B and C POPF require interventional therapy, such as endoscopic or percutaneous procedures [66,68,69], and a few patients should receive re-operation [70][71][72].

Non-surgical treatment
Sufficient drainage is the basis of treatment for POPF. In most cases, peritoneal fluid collection caused by pancreatic fistula could be effectively managed through routine abdominal drainage [6,29], percutaneous peritoneal drainage under CT or ultrasound guidance [29,33,47,73], and internal drainage using endoscopic ultrasonography by placing pancreatic duct stents [17,29]. With sufficient drainage combined with nutritional support [73], antibiotic therapy [74], and inhibition of pancreatic secretion, most patients would recover uneventfully from POPF.
Postoperative feeding time and route of nutritional treatment have always been the key issues of postoperative nutritional support for pancreatic surgery. Premature feeding or enteral nutrition may promote the secretion of pancreatic juice, which may hamper the healing of pancreatic fistula. On the other hand, fasting or long-term parenteral nutrition can also lead to dystrophy, healing difficulties, and parenteral nutrition-related complications [75,76]. In 2006, a retrospective study conducted by Pannegeon et al. [76] confirmed that POPF could be relieved by conservative treatment including enteral and parenteral nutrition and abdominal drainage. As fasting is sometimes considered necessary to suppress the secretion of pancreatic juice in patients with POPF, a multi-center RCT in 2015 proved that food intake did not aggravate POPF, which indicated that there may be no need to avoid oral dietary intake in patients with POPF [75].

Surgical treatment
Surgical treatment should be considered for patients with POPF who fail to respond to conservative treatment. After non-surgical treatment, patients that still have persistent pancreatic fistula, abdominal bleeding, abdominal abscess, sepsis, and other critical situations require further surgical treatment [24,29,39,47,73].
Reoperation approach includes placement of drainage in the vicinity of pancreatic stump or anastomosis site (the pancreatic stump could be further resected and sutured again) [73,74], anastomosis of pancreatic stump and digestive tract, and removing a part of pancreatic tissue or even the whole pancreas. In case of ischemic necrosis of other organs, combined resection of vessels and organs is even required [29].

Summary
With the development of minimally invasive surgery, the surgical technique of pancreatic body and tail resection is also constantly improving. Minimally invasive radical resection is gradually being popularized and has achieved good results in clinical practice. POPF is still the most common and dangerous complication of pancreatic body and tail resection. Experts and scholars are attempting to standardize the definition of pancreatic fistula in order to diagnose, prevent, and treat pancreatic fistula more reasonably. They are continuously studying the risk factors of pancreatic fistula so as to avoid its occurrence at an early stage. Many new approaches (e.g., pancreatic disconnection and stump closure) and materials are being developed and utilized to prevent pancreatic fistula.
Prospective and randomized controlled studies on body and tail pancreatectomy are relatively few, and this should be the direction of future research and development.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.