For pancreaticoduodenectomy, a fistula risk score has been recently developed that has been shown to be highly predictive of POPF. This score assigns points based on gland texture, gland pathology, duct diameter, and intraoperative blood loss. In general, high blood loss, soft gland texture, and smaller duct diameter confer increased risk of POPF, whereas pancreatic adenocarcinoma and pancreatitis as the indication for pancreaticoduodenectomy confer protection for the development of pancreatic fistula versus other diagnoses. Also of note, higher fistula risk scores correlated with greater incidence of clinically relevant (ISGPF grade B or C) fistula [9]. The adaptation of this risk score to patients undergoing distal pancreatectomy is yet to be validated; however, at least one published study indicates that this scoring system may have limitations in the setting of distal pancreatectomy. In that study, risk factors for pancreatic fistula after stapled gland transection in patients undergoing distal pancreatectomy were examined, and in a multivariate analysis, only the presence of diabetes and the use of a 4.1-mm staple cartridge were associated with increased risk of pancreatic fistula formation [10].

Some retrospective data have supported the conclusions drawn with regard to risk for pancreatic fistula after pancreaticoduodenectomy in the setting of distal pancreatectomy. A retrospective case-matched analysis looked at histopathologic features of patients with fistula and matched patient controls and found that gland fat content, smaller main duct size, and the lack of stigmata of chronic pancreatitis or interlobular fibrosis were correlated with increased risk of POPF. This study included 9 patients who underwent pancreaticoduodenectomy as well as 16 patients who underwent distal pancreatectomy. A score was created using these and other characteristics, which had 92 % sensitivity and 84 % specificity for the postoperative development of pancreatic fistula [11].

The somatostatin analog octreotide has been the subject of numerous investigations in pancreatic surgery and specifically in the prevention of the formation of pancreatic fistula following pancreaticoduodenectomy or distal pancreatectomy. The biologic plausibility of the effectiveness of octreotide is significant; by decreasing pancreatic secretions, pressure gradients across the pancreatic ductal anastomosis or closure would be decreased, thereby resulting in decreased fistula rates. Alternatively, where a fistula has already formed, octreotide might have the potential to convert a high-output fistula into a low-output fistula, decreasing the likelihood of complications and increasing the chances of a spontaneous closure. However, most studies to date have failed to definitively identify a role for octreotide in pancreatic surgery. Furthermore, the biologic rationale, while emotionally appealing, may not stand up to scientific scrutiny; a 2013 single-institution trial measured the effect of octreotide in patients who underwent a pancreaticoduodenectomy by directly measuring exocrine output using intraductal pancreatic catheters and failed to demonstrate any significant difference between octreotide and placebo in the volume of pancreatic exocrine secretion [12].

Recently, a multicenter randomized controlled trial was undertaken to answer the question of whether octreotide administration was of benefit in pancreatic surgery. Enrolling 230 patients with slightly more than half-randomized to the octreotide group, the study failed to demonstrate any overall benefit in octreotide administration. While a subgroup analysis suggested some benefit for patients with small duct diameters, the study overall was significantly weakened by a significant increase in the incidence of intraductal fibrin sealant administration in the octreotide group [13].

Stenting of the main pancreatic duct at the ampulla has been investigated as a method to prevent postoperative fistula formation after distal pancreatectomy. This again has a clear anatomic rationale, as decompression of the pancreatic tree via drainage across the ampulla would have the effect of decreasing pressure against the pancreatic stump, thereby reducing the likelihood of leakage from the resection margin. Retrospective data from several centers initially demonstrated some success with the technique. One small series published in 2008 noted a 20 % incidence of mild pancreatitis but no instances of pancreatic leak among ten patients who underwent distal pancreatectomy with prior endoscopic placement of transampullary stent [14]. A second retrospective series involved the intraoperative placement of transampullary stents. In this series, the surgeons identified the transected duct at the resection margin of the distal pancreatectomy and subsequently advanced a pediatric feeding tube into the duodenum. The transected end of the duct was then ligated. The authors were able to demonstrate an association of intraoperative stent placement with decreased pancreatic fistula rates as well as decreased overall length of stay [15].

However, in 2012, a randomized prospective trial was performed in which patients were assigned either to distal pancreatectomy alone or to distal pancreatectomy with prior transpapillary stent placement. That trial failed to show a benefit of preoperative stenting of the pancreatic duct and in fact demonstrated a trend toward a significantly increased rate of pancreatic fistula among patients with preoperative stent placement [16]. An example that highlights the significant difficulty in obtaining robust best practice standards in this area comes from a similar controversy in pancreaticoduodenectomy. Although the operations are significantly different, prior retrospective examination of patients undergoing pancreaticoduodenectomy with intraoperative stent placement demonstrated no benefit for the prevention of POPF [17]. However, a randomized trial performed around the same time actually demonstrated a benefit to intraoperative pancreatic stenting [18].

The management of the pancreatic stump created during a distal pancreatectomy has also been the subject of some controversy. Historically, techniques of transection of the pancreatic stump included sharp division and oversewing of the transected surface. With the advent of staplers, controversy has arisen with regard to their use in transection of the pancreatic body. Some authors initially suggested that hand-sewn closures had lower rates of fistula, whereas others demonstrated superior results with stapler use [19]. Recently, a multivariate analysis identified increased thickness of the pancreatic body as a risk factor for failure with stapled transections. Additionally, use of a double-row (as opposed to a triple-row) stapler load was associated with increased risk of fistula [20].

Anatomic techniques have also been widely investigated. A recently published randomized controlled trial compared stump reinforcement with fibrin glue and a falciform patch to no reinforcement among patients undergoing distal pancreatectomy with stapled or hand-sutured stump closure techniques, and found identical rates of pancreatic fistula among the two groups [21].

Another group of techniques described in the literature include the creation of anastomoses between the pancreatic stump and either the bowel or the stomach. One prospective case series described 21 patients undergoing distal pancreatectomy with the creation of pancreaticogastrostomy, and the authors were able to report a 0 % rate of grade B or C pancreatic fistula [22]. Another group in a retrospective review demonstrated a statistically significant elimination in the number of pancreatic fistula when a roux-en-Y limb was brought up to provide distal drainage to the transected pancreatic stump after distal pancreatectomy [23]. Indeed, numerous studies have reported somewhat favorable results with creation of an anastomosis at the distal pancreatic stump [24]. A large series by Kleef et al. [25] examined 302 patients undergoing distal pancreatectomy with an overall fistula rate of 12 %. Data were gathered prospectively, and four main techniques of pancreatic stump management were described: (1) pancreaticojejunostomy; (2) seromuscular patch; (3) suture of the duct with polydioxanone (PDS) stitch followed by parenchymal closure with PDS suture, with or without collagen patch; (4) closure with a stapler, primarily with a vascular load. In this series, a stapled anastomosis was associated with a higher rate of fistula formation. As this was a retrospective, single-center review, certain subgroup analyses such as a strict comparison of patient characteristics between the various closure techniques were not reported [25]. Additionally, the use of nonvascular stapler loads, which the authors suggest were occasionally used, may be quite significant. One single-center study suggested that the rate of pancreatic fistula was much lower when a vascular (2.5 mm) cartridge was utilized instead of a standard cartridge or a hand-sewn technique [26]. More recently, a large multicenter trial was conducted to answer the question of whether a stapled or hand-sewn technique prevented POPF. The DISPACT trial randomized patients into a stapled or hand-sewn closure of the pancreatic stump. The primary endpoints included combined mortality and/or the detection of pancreatic fistula prior to postoperative day #7. Secondary endpoints included detection of pancreatic fistula up until postoperative day #30. No difference in fistula rates was described between the two groups. Additionally, outcomes among a range of clinical factors were similar. A post hoc analysis was conducted within the trial to determine the factors associated with the development of pancreatic fistula and did not reveal any factor to be causative in a multivariate analysis [27].

Ligation of the main pancreatic duct, where technically feasible, has been reported to dramatically reduce the incidence of pancreatic duct leak in some studies, with conflicting reports in others. The main limitation of data addressing this technical point is that it is limited to single-institution retrospective studies. Some authors have been able to demonstrate a reduction in the rate of pancreatic fistula from greater than 30 % to less than 10 % and that the performance of duct ligation was a significant negative predictor of pancreatic fistula by multivariate analysis [28]. Notably, in that study, which included an overall pancreatic leak rate of approximately 20 %, other factors including pancreatic pathology, hand-sewn or stapled closure, octreotide use, blood transfusion, and operating time, among other factors, were all demonstrated to be unrelated to the postoperative development of pancreatic leak in a multivariate analysis. A recent retrospective review of 704 patients undergoing distal pancreatectomy at a single institution was not able to detect a significant effect of duct ligation on the prevention of pancreatic leak, and in fact detected a trend toward increased clinically significant leak rate when duct ligation was performed. Of note, however, duct ligation was employed selectively at this institution, and thus may have been reserved for those cases with large duct diameters or other intraoperative findings that raised concern for increased likelihood of duct leak [3].

There has been new interest in managing the transected pancreatic stump by reinforcing the stump with a mesh closure. Early retrospective data from single centers suggested that use of an absorbable mesh to reinforce the staple line of the transected pancreas reduced the rate of stump leak [29]. This method has been investigated with a randomized, single-blinded clinical trial with a total enrollment of 100 patients. Reinforcement of the distal pancreatectomy resection margin with mesh reduced the rate of clinically significant (ISGPF B and C) pancreatic fistula from 20 % to less than 2 % [30]. One potential disadvantage of this technique relates to expense, as the placement of a mesh significantly increases the cost of operation. However, a recent cost analysis suggested that patients who received mesh placement during distal pancreatectomy had overall lower hospital charges and decreased length of stay versus patients who underwent distal pancreatectomy without mesh placement [31].

Despite advances in laparoscopic and robotic approaches, the vast majority of distal pancreatectomies continue to be performed via an open approach. Recent retrospective data have demonstrated that minimally invasive distal pancreatectomy is associated with decreased blood loss and shorter hospital stays than open pancreatectomy [32]. A large recent study utilizing the Nationwide Inpatient Sample database suggested, first, that the minimally invasive approach is becoming more widely utilized, increasing from 2.4 to 7.3 % over a study period from 1998 to 2009. Second, that study reported that the minimally invasive approach was associated with decreased length of stay as well as decreased incidence of infectious complications, bleeding complications, and blood transfusions [33]. This population-based study echoes conclusions drawn by a large multi-institutional study performed several years previously. Drawing on a combined patient sample of 667 patients, with 24 % initially attempted laparoscopically, the authors were able to demonstrate lower overall complication rate, decreased blood loss, and shorter hospital stays among patients undergoing laparoscopic approach via a multivariate analysis. Notably, there was no significant difference in the pancreatic leak rate between the open and laparoscopic approaches, although there was a nonsignificant trend favoring the laparoscopic approach [34].