The occurrence of a pancreatic fistula is highly dependent on the type of surgical procedure performed and the underlying pancreatic pathology. Soft pancreatic tissue texture without pre-existing fibrosis is regarded as a risk factor for fistula development in all pancreatic procedures.

Given the frequency of pancreatic fistulas following pancreatic resection, extensive research has been employed to prevent the occurrence of pancreatic fistula. The strategies include pharmacologic manipulation, modifications and refinements in surgical technique regarding pancreatic anastomosis, pancreatic anastomotic stents, and perianastomotic drainage post pancreatic resection.

Octreotide, a synthetic somatostatin analogue inhibits pancreatic exocrine secretion. The use of octreotide and its analogues to prevent postoperative fistula is an approach which has been used since the 1990s [29, 30]. Despite 20 years of clinical use and evaluation in numerous studies, a recent Cochrane meta-analysis concluded that evidence is still lacking to give clear guidelines [31]. While early randomized controlled trials (RCTs) favored the use of octreotide and showed a 50 % reduction of fistula rates, these findings were not confirmed in later studies [30, 32, 33]. From these results, it was concluded and supported by the Cochrane review that routine use of octreotide was not indicated, but should be used in a risk-dependent manner in a presumed “critical” anastomoses due to soft pancreatic tissue texture. Although the overall fistula rates have been reduced, somatostatin analogues failed to reduce the incidence of clinically relevant (grade B/C) fistula or re-operation rates and mortality. Finally, postoperative octreotide administration for postoperative fistula has failed to show any improvement in the rate of fistula closure [34]. Despite the lack of effect on fistula closure rate, the octreotide may help lower fistula output and make fistula control easier. Recently, a single-center, randomized, double-blind trial was published showing pasireotide, a new somatostatin analogue, decreased the rate of clinically significant postoperative pancreatic fistula, leak, or abscess [35]. While the initial results are promising, further studies need to be conducted to prove whether pasireotide is beneficial.

Modifications in surgical technique to prevent pancreatic fistula have been evaluated for decades with conflicting results. Following a pancreaticoduodenectomy, there has been an ongoing debate of whether a pancreaticogastrostomy or pancreaticojejunostomy has a lower postoperative fistula rate. Within each of these techniques, several different technical modifications, including single or double layer sutures, invagination and purse-string sutures have been compared with regard to the surgical complications and especially postoperative fistula frequency. Unfortunately, there have been no level 1 evidence-supported techniques that have been universally adopted. Following distal pancreatectomy, there are numerous studies comparing sutures, staplers, patches, fibrin glue, and sealants to handle the distal stump of the pancreatic remnant that have been reported [13, 14], and no convincing evidence exists to support the superiority of any one technique.

The placement of pancreatic duct stents and the potential role in prevention of postoperative pancreatic fistula has been investigated for both right and left pancreatic resections. The principle of internal drainage of the pancreatic duct following pancreaticoduodenectomy to achieve a diversion of the pancreatic secretion from the suture site has been hypothesized; however, the stent may also cause problems via irritation of the duct and the suture lines as well as the obstruction or migration. In available studies, the outcome shows a great deal of variability. Earlier studies demonstrated a beneficial effect of anastomotic stenting in lowering the postoperative pancreatic fistula rate [36, 37]. A randomized trial by Poon et al. among 120 patients, which used long stents across the pancreaticjejunostomy anastomosis and drained externally, showed that the stented group had a significantly lower pancreatic fistula rate compared to the nonstented group (6.7 vs. 20 %, respectively) [37]. Despite these encouraging results, the technique has not been universally adopted. In contrast, the largest randomized study, published in 2006 by Winter et al., which used short internalized stents (6 cm long plastic pediatric feeding tube) included 234 patients who underwent pancreaticoduodenectomy with stent ( n = 115) or without stent ( n = 119) placement into the pancreatic duct [38]. Winter showed an overall fistula rate of 7.6 % (no stent) vs. 11.3 % (stent), concluding no benefit for stenting of the pancreatic duct. The most recent study, published in 2012 by Sachs et al. where 59/444 patients had an intraoperatively pancreaticojejunal stent placed actually had greater rates of critically relevant postoperative pancreatic fistula, major complications, greater length of stay, and total costs [39].

There are very few studies analyzing the effect of preoperative or intraoperative stents in the distal pancreatectomy setting. In a 23-patient collective, Fischer et al. described a prophylactic intraoperative transampullary stent placement as an open surgical procedure that resulted in a significant reduction in postoperative pancreatic fistula rates [40]. However, Okamoto et al. observed a stent-related morbidity of 57 %, including pancreatitis and stent obstruction [41]. Reider et al. had no postoperative pancreatic fistula in stented patients; however, in this study a sphincterotomy was performed in addition to the placement of preoperative stents prior to distal pancreatectomy [42].

The routine use of intraperitoneal drains following elective pancreatic surgery remains an area of debate regarding whether drains prevent or exacerbate pancreatic fistulas. The first randomized trial to investigate the impact of intraperitoneal drain use reported by Conlon et al. randomized 179 patients following pancreaticoduodenectomy and distal pancreatectomy to intraperitoneal drain placement or surgery without drains [43]. This study demonstrated that patients in the drainage group were more likely to have an intra-abdominal abscess, collection, or fistula compared to patients without drains. More recently, Bassi et al. randomized patients who were at low risk for leak based on drain amylase level < 5000 U/L on postoperative day 1 to early drain removal (postoperative day 3) or late drain removal (≥ postoperative day 5) [44]. The trial reported a significant reduction in pancreatic fistulas (using the ISGPF definition) in early drain removal group versus the late drain removal group 1.8–26.3 %. While these studies show that drains may not be necessary, most surgeons routinely still use postoperative drains and have not changed their clinical practice. Recently, a randomized multicenter trial was conducted to further evaluate the necessity of drains after pancreaticoduodenectomy [45]. There were no differences between the drain and no-drain cohorts in demographics, comorbidities, pathology, pancreatic duct size, pancreas texture, or operative technique. Pancreaticoduodenectomy without intraperitoneal drainage was associated with an increase in the number of patients with complications, number of complications per patient, and the severity of complications. The no-drain cohort also had a higher incidence of gastroparesis, intra-abdominal fluid collection, intra-abdominal abscess, and severe diarrhea. Furthermore, patients in the no-drain group more often required postoperative percutaneous drains and had a prolonged hospital stay. The Data Safety Monitoring Board stopped the study early because of the increased mortality from 3 to 12 % in the patients undergoing pancreaticoduodenectomy without intraperitoneal drainage. Thus, most surgeons feel that this level 1 evidence provides strong support for routine drainage following pancreaticoduodenectomy. This same study continues to determine the necessity of drains for patients after distal pancreatectomy.

Multiple studies have demonstrated that patients with pancreatic leak have a significant increase in secondary complications compared to patients without leak [8, 46]. Commonly observed complications are mainly caused by undrained infected pancreatic fluid collections. Pancreatic fluid is an enzymatically active and aggressive substance that may cause erosion of the surrounding tissue, organs, and blood vessels. This can lead into leakage from other adjacent anastomoses or bowel (particularly in patients with pancreatic necrosis) causing a biliary, gastric, or enteric leak.

Postoperative hemorrhage associated with a pancreatic leak is one for the most dreaded complications following major pancreatic resections. The pancreatic enzymes in combination with infection can cause erosion of the gastroduodenal artery or splenic artery stump or from an arterial pseudoaneurysm resulting in significant bleeding requiring immediate therapy. This complication usually occurs after the first week after the surgery and in most cases with what appears to be adequate drainage of the pancreatic leak. Management is guided by the patient’s clinical status and hemodynamic stability. In general, most patients should be approached via angiographic embolization or arterial stenting to provide the best outcomes [47]. Patients who are hemodynamically unstable may require operative re-exploration and packing and then angiographic control. A high index of suspicion should be maintained because postoperative hemorrhage is associated with significant risk.

The occurrence of pancreatic fluid collections due to a pancreatic leak is also a potential cause of ongoing abdominal sepsis that can lead to generalized systematic organ failure. Percutaneous drainage of the fluid collections by interventional radiology and broad spectrum antibiotic therapy are as important as supportive ICU therapy in these patients.

An important aspect of the pancreatic fistula complications is the economic impact from the prolonged treatment. The longer duration of hospital stay is an important factor that increases treatment costs. The average hospital stay in uncomplicated resections is usually 6–8 days, but can increase to 25–40 days in cases of fistula development, especially with type B or C fistula [46]. The associated treatment costs in these patients are 4–5 times higher than in patients without fistulas, highlights the socio-economic dimension of the health care system [8].

Regardless of the cause or the location of the pancreatic fistula, the steps required for treatment of a clinically relevant pancreatic fistula are similar. First, stabilization of patients and medical optimization are the crucial steps. Drainage of collections and insuring operatively placed drains are adequately controlling the fistula output to control sepsis that is mandatory. In cases with sepsis or high output fistulas, the patient is made “nil per os” (NPO) and parenteral nutrition is considered necessary. Only then should the nature of pancreatic duct injury be investigated and definitive management of the fistula be addressed.