Pancreaticojejunostomy versus pancreaticogastrostomy after pancreaticoduodenectomy to prevent post-operative pancreatic fistula, a dissonance between evidence and practice

Leakage of pancreatic enzymes leading to either formation of abdominal collection or pancreatic fi stula is one the most feared complications after pancreaticoduodenectomy. Owing to high morbidity and cost related to pancreatic fi stula, multiple interventions including various types of pancreaticoenteric anastomosis have been proposed to prevent this complication. Despite some randomized controlled trials and meta-analyses favoring pancreaticogastrostomy over pancreaticojejunostomy, clinical practice has not witnessed any change in preference of individual surgeons. One of the underlying facts is that there are various ways of doing pancreatic anastomosis and trials have compared only specifi c techniques while a few novel techniques that have been reported to have very low pancreatic fi stula risk have never been compared in randomized controlled trials comparing pancreaticogastrostomy versus pancreaticojejunostomy. Moreover individual surgeons’ comfort and training also matters, and in many instances same results are not reproduced as reported for primary center where technique was developed. So though a good number of randomized controlled trials have been conducted to compare pancreaticogastrostomy with pancreaticojejunostomy, variations in techniques of performing anastomosis limit external validity as well as pooling the data for meta-analysis. Furthermore subgroup of patients with soft pancreas, who are at high risk of pancreatic leak, should be looked at separately for potential benefi t of type of pancreatic anastomosis. Review Article Pancreaticojejunostomy versus pancreaticogastrostomy after pancreaticoduodenectomy to prevent post-operative pancreatic fi stula, a dissonance between evidence and practice Noman Shahzad1*, Tabish Umer Chawla2, Saleema Begum3 and Fareed Ahmed Shaikh4 1Consultant General Surgeon, Aga Khan University Hospital Karachi, Pakistan 2Section Head General Surgery, Aga Khan University Hospital Karachi, Pakistan 3Fellow General Surgery, Aga Khan University Hospital Karachi, Pakistan 4Fellow Vascular Surgery, Aga Khan University Hospital Karachi, Pakistan Received: 22 February, 2018 Accepted: 03 March, 2018 Published: 06 March, 2018 *Corresponding authors: Dr. Noman Shahzad, Consultant General Surgeon, Aga Khan University Hospital Karachi, Offi ce of General Surgery, Link Building, Aga Khan University Hospital Karachi. Postal Code: 74800, Pakistan, Tel: +923008833313 E-mail:


Introduction
Pancreaticoduodenectomy (PD) remains the only curative option for resectable pancreatic head, ampullary, duodenal and distal common bile duct tumors. Despite improvements in post-operative care and advancement in surgical techniques, morbidity related to this operation remains very high. According to recent report by St-Germain AT et al. up to 74% of patients suffer from at least one complication related to this complex surgical procedure [1]. Leakage of pancreatic enzymes leading to either formation of abdominal collection or pancreatic fi stula is one the most feared complications. Incidence of post-operative pancreatic fi stula (POPF) after PD is reported to be from 11% to 47.7% in various reports [2,3]. This wide variation in occurrence of POPF is partly due to variability in defi nition of fi stula particularly in older studies. Criteria to label pancreatic fi stula was standardized by international study group on pancreatic fi stula (ISGPF) in 2005 [4].
Furthermore due to high morbidity and cost related to pancreatic fi stula [5], multiple interventions have been investigated to prevent this complication [6]. These include pharmacological interventions such as role of peri-operative octreotides administration, adjuncts to surgical anastomosis such as stenting of anastomosis or use of sealants, surgical techniques and site of pancreatico-enteric anastomosis.
Of these, comparison of pancreaticogastrostomy with pancreaticojejunostomy is the most studies area. To the best of our knowledge ten randomized controlled trials have been conducted to date to fi nd out better site of performing pancreatic anastomosis. Three of these trials concluded that pancreaticogastrostomy is superior to pancreaticojejunostomy post-operative pancreatic fistula, a dissonance between evidence and practice. Arch Clin Gastroenterol 4(1): 008-011. DOI: http://doi.org/10.17352/2455-2283.000048 to prevent POPF [7][8][9], while others failed to detect any signifi cant difference. Pooling of data in reported meta-analysis has also not been able to reach a defi nitive conclusion. Of the two most recent meta-analyses reported, one has concluded that pancreaticogastrostomy is superior pancreaticojejunostomy to prevent POPF [10] while the other concludes that there is no statistically signifi cant difference [11].
Other than site of pancreatic anastomosis, details of surgical technique employed for anastomosis and individual surgeon variations are the factors to be kept in mind while looking at evidence related to pancreatic fi stula. Moreover surgical approach is altered in many instances of high risk features; its value in pancreatic anastomosis needs to be explored. To the best of our knowledge, there are ten randomized controlled trials conducted to date to compare pancreaticogastrostomy versus pancreaticojejunostomy. Three of these trials found that occurrence of pancreatic fi stula was signifi cantly lower in pancreaticogastrostomy group [7][8][9]. These trials had failed to detect any difference in the two groups [11], but ran-dom effect model was used to analyze the results as opposed to former meta-analysis. However there is no trial or metaanalysis published as yet that reported superiority of pancreaticojejunostomy over pancreaticogastrostomy.

Differences in surgical techniques
Individual Surgeon Variations: There are several ways of doing pancreatic anastomosis and employing one way of doing anastomosis as opposed to the other depends upon comfort and training of operating surgeon in addition to other factors. Adopting and mastering another way of doing the same task when surgeon is comfortable with one way is not always easy and may not reproduce the same results as proposed by other surgeons. This is why same technique has different rates of pancreatic fi stula reported from different centers [13].
Other than conventional technique, there are reported improvisations with promising results, but not all of these have been studied in randomized controlled trials comparing pancreaticogastrostomy with pancreaticojejunostomy.
Pancreaticojejunostomy: Conventionally pancreaticojejunostomy is performed as end to side, double layer, duct to mucosa anastomosis in which inner layer incorporates full thickness jejunal wall to pancreatic duct and outer layer as seromuscular jejunal stitch to pancreatic tissue. Reported leak rate after conventional technique is 6-22% [14]. Invagination of pancreatic tissue with or without duct to mucosa stitches has been studied with promising results. Invagination with duct to mucosa stitches is reported to have rate of CR-POPF as low as 3.3% [15]. Binding pancreaticojejunostomy as described by Peng et al incorporates destruction of 3 cm jejunal mucosa by applying 10% carbolic acid followed by rinsing with 75% alcohol and normal saline. After doing pancreaticojejunal anastomosis an absorbable ligature is looped around the jejunum, with the invaginated pancreas inside. Randomized controlled trial comparing binding pancreaticojejunostomy with conventional technique found signifi cantly lower fi stula rate for binding technique. It reported no pancreatic fi stula in 106 patients randomized to binding technique group [16]. This technique is not compared to pancreaticojejunostomy in any of the randomized controlled trials. Moreover similar results could not be obtained for this technique at other centres. Maggiori et al., in their study reported no decrease in pancreatic fi stula, rather risk of haemorrhage was increased [17].
Isolated loop pancreaticojejunostomy has also been compared with pancreaticogastrostomy in randomized controlled trial and no signifi cant difference was found in pancreatic fi stula rate [18].
Pancreaticogastrostomy: Conventionally pancreaticogastrostomy is performed as invaginated double layer anastomosis to posterior wall of stomach. Fernandez et al., reported doing pancreaticogastrostomy with gastric partition in which they made pancreaticogastric anastomosis to partitioned part of stomach. They compared it with conventional pancreaticojejunostomy in a randomized controlled trial and demonstrated that this technique was signifi cantly superior to pancreaticojejunostomy in reducing pancreatic fi stula risk [7]. It has been proposed that lack of enterokinase and acidic environment in stomach inactivates pancreatic enzymes, which along with good blood supply of stomach may have role to play in reducing risk of anastomotic leak [19]. While potential of anastomotic leak is reduced by pancreaticogastrostomy, long term exocrine and endocrine functions are compromised more in these patients as compared to those who underwent pancreaticojejunostomy [20]. Furthermore risk of digestive tract bleeding is also more after pancreaticogastrostomy, though management of GI bleed is easy via upper gastrointestinal endoscopy should bleeding occur [21].
Subgroup at high risk of leakage: In addition to postoperative care and surgical technique, certain patient and disease related factors predispose patients to high risk of POPF development [22]. Soft texture of pancreas is an established risk factor for POPF [23]. There are only a few randomized controlled trials that have been conducted on or have reported separate subgroup analysis for this select subgroup of patients. Bassi et al., reported on difference in fi stula rate after pancreaticogastrostomy versus pancreaticojejunostomy for patients with soft pancrease [24]. Contrary to that, subgroup of patients with soft pancreas in randomized controlled trial by Topal et al., demonstrated that pancreaticogastrostomy was superior to pancreatiocojejunostomy for postoperative pancreatic fi stula [9]. There has been no metaanalysis to date to compare pancreaticogastrostomy versus pancreaticojejunostomy in patients with intra-operative soft texture of pancreas which needs to be addressed via pooled data analysis.

Conclusion
Though a good number of randomized controlled trials have been conducted to compare pancreaticogastrostomy versus pancreaticojejunostomy, variations in techniques of performing anastomosis limit external validity. Furthermore this issue of variability in surgical technique across randomized controlled trials should be taken care of before pooling the data for meta-analysis. Moreover subgroup of patients with soft pancreas who are at high risk of pancreatic leak, should be looked at separately for potential benefi t of site of anastomosis.
In addition to that, other than statistical evidence, to change practice where learning of a new skill is required, many other factors including training, learning curve and required facilities have to be accounted for.