Preventing Pancreatic Fistula After Pancreaticoduodenectomy
Preventing Pancreatic Fistula After Pancreaticoduodenectomy
The results of this meta-analysis show that PG is superior to PJ for reconstruction after PD. The results indicate that PG is associated with significantly lower postoperative pancreatic and biliary fistula rates and a shorter length of hospital stay than PJ. However, these 2 procedures were not significantly different in terms of delayed gastric emptying, overall morbidity, or mortality.
Pancreatic anastomosis after PD is an important aspect of pancreatic surgery. POPF formation is the most important cause of morbidity and mortality after PD. Despite recent improvements, this complication still occurs in 30% of cases. Other complications after PD such as biliary fistula formation and delayed gastric emptying are also concerning. Previous studies reported contradictory results regarding the impact of PG versus PJ on the postoperative fistula rate. Previous retrospectives studies, prospective studies, randomized controlled trials, and 5 meta-analyses have reported on the technical aspects influencing the pancreatic fistula rate. Pharmacological treatment has also been used to help reduce the pancreatic fistula rate.
Numerous PJ anastomotic techniques have been described, using end-to-end or side-to-end anastomoses, with or without invagination of the pancreas into the digestive tract in a single layer or double layers. In PG, the remnant pancreas is anastomosed to the posterior wall of the stomach, with or without invagination of the pancreas. This procedure is easy to perform because the posterior wall of the stomach can be mobilized toward the pancreas. Several explanations regarding the protective role of PG have been described in the literature. Some authors have suggested that pancreatic secretions may be less corrosive to the stomach after PG than to the digestive tract after PJ, because the acidity of gastric secretions inactivates the pancreatic enzymes. In PJ, the pancreatic enzymes are activated by alkaline biliary and enteric secretion. Furthermore, gastric and pancreatic secretion is easily diverted with a nasogastric tube after PG, and PG reconstruction may divert potential pancreatic fistulas away from major blood vessels. In PJ, complex fistulas may form with leakage of biliary and pancreatic secretion.
Previous meta-analyses have also investigated pancreatic fistula rates after PD. Only Shen et al found no significant difference in the pancreatic fistula rate between PG and PJ. They included only 4 randomized controlled trials in their analysis, which may have been an insufficient number of patients to reach definitive conclusions. Ma et al found that PG was associated with a lower pancreatic fistula rate than PJ. However, their data had significant heterogeneity because they included both randomized controlled trials and prospective trials in their analysis. He et al and Wente et al included both randomized controlled trials and observational studies in their analyses and did not find superiority of either PG or PJ. More recently, the meta-analysis by Yang et al could not determine the best method of reconstruction among PG, PJ, intraduct ligation of the pancreatic duct, duct-to-mucosa PJ, and binding PJ.
Other factors also influence the pancreatic fistula rate after PD. For example, a fatty pancreas and a pancreatic duct size of less than 3 mm are associated with an increased risk of POPF. Other patient factors also affect the risk of pancreatic fluid leakage, including age, preoperative jaundice, body mass index, and cardiovascular comorbidities.
Our meta-analysis has some limitations. First, the types of intervention and the indications for surgery are heterogeneous among published studies. For example, Duffas et al, Bassi et al, Yeo et al, and Topal et al reported outcomes after PD and Fernandez-Cruz et al, Topal et al, and Wellner et al reported outcomes after pylorus-preserving PD. These different interventions may lead to different complications. Second, the definition of pancreatic fistula varied among studies, and the ISGPF definition was used in only 4 of the 7 studies included in our analysis. Finally, it would have been useful to consider the reason for performing PD in the analysis, but this was not possible because of the limited information available.
In the study by Yeo et al, one patient who died was excluded from the study because his death was judged to be caused by pulmonary embolism and multiorgan failure. We excluded this study from the meta-analysis of postoperative mortality but included it in the other groups to determine whether this would alter the conclusions and found that it did not.
PG seems to be associated with a shorter length of hospital stay than PJ. This information should be interpreted with care because the definition of length of hospital stay varied among studies (eg, only postoperative days were reported in the study by Yeo et al). There was also a higher rate of rehospitalization in the PG group than in the PJ group in the study by Figueras et al, but this difference was not statistically significant.
Discussion
The results of this meta-analysis show that PG is superior to PJ for reconstruction after PD. The results indicate that PG is associated with significantly lower postoperative pancreatic and biliary fistula rates and a shorter length of hospital stay than PJ. However, these 2 procedures were not significantly different in terms of delayed gastric emptying, overall morbidity, or mortality.
Pancreatic anastomosis after PD is an important aspect of pancreatic surgery. POPF formation is the most important cause of morbidity and mortality after PD. Despite recent improvements, this complication still occurs in 30% of cases. Other complications after PD such as biliary fistula formation and delayed gastric emptying are also concerning. Previous studies reported contradictory results regarding the impact of PG versus PJ on the postoperative fistula rate. Previous retrospectives studies, prospective studies, randomized controlled trials, and 5 meta-analyses have reported on the technical aspects influencing the pancreatic fistula rate. Pharmacological treatment has also been used to help reduce the pancreatic fistula rate.
Numerous PJ anastomotic techniques have been described, using end-to-end or side-to-end anastomoses, with or without invagination of the pancreas into the digestive tract in a single layer or double layers. In PG, the remnant pancreas is anastomosed to the posterior wall of the stomach, with or without invagination of the pancreas. This procedure is easy to perform because the posterior wall of the stomach can be mobilized toward the pancreas. Several explanations regarding the protective role of PG have been described in the literature. Some authors have suggested that pancreatic secretions may be less corrosive to the stomach after PG than to the digestive tract after PJ, because the acidity of gastric secretions inactivates the pancreatic enzymes. In PJ, the pancreatic enzymes are activated by alkaline biliary and enteric secretion. Furthermore, gastric and pancreatic secretion is easily diverted with a nasogastric tube after PG, and PG reconstruction may divert potential pancreatic fistulas away from major blood vessels. In PJ, complex fistulas may form with leakage of biliary and pancreatic secretion.
Previous meta-analyses have also investigated pancreatic fistula rates after PD. Only Shen et al found no significant difference in the pancreatic fistula rate between PG and PJ. They included only 4 randomized controlled trials in their analysis, which may have been an insufficient number of patients to reach definitive conclusions. Ma et al found that PG was associated with a lower pancreatic fistula rate than PJ. However, their data had significant heterogeneity because they included both randomized controlled trials and prospective trials in their analysis. He et al and Wente et al included both randomized controlled trials and observational studies in their analyses and did not find superiority of either PG or PJ. More recently, the meta-analysis by Yang et al could not determine the best method of reconstruction among PG, PJ, intraduct ligation of the pancreatic duct, duct-to-mucosa PJ, and binding PJ.
Other factors also influence the pancreatic fistula rate after PD. For example, a fatty pancreas and a pancreatic duct size of less than 3 mm are associated with an increased risk of POPF. Other patient factors also affect the risk of pancreatic fluid leakage, including age, preoperative jaundice, body mass index, and cardiovascular comorbidities.
Our meta-analysis has some limitations. First, the types of intervention and the indications for surgery are heterogeneous among published studies. For example, Duffas et al, Bassi et al, Yeo et al, and Topal et al reported outcomes after PD and Fernandez-Cruz et al, Topal et al, and Wellner et al reported outcomes after pylorus-preserving PD. These different interventions may lead to different complications. Second, the definition of pancreatic fistula varied among studies, and the ISGPF definition was used in only 4 of the 7 studies included in our analysis. Finally, it would have been useful to consider the reason for performing PD in the analysis, but this was not possible because of the limited information available.
In the study by Yeo et al, one patient who died was excluded from the study because his death was judged to be caused by pulmonary embolism and multiorgan failure. We excluded this study from the meta-analysis of postoperative mortality but included it in the other groups to determine whether this would alter the conclusions and found that it did not.
PG seems to be associated with a shorter length of hospital stay than PJ. This information should be interpreted with care because the definition of length of hospital stay varied among studies (eg, only postoperative days were reported in the study by Yeo et al). There was also a higher rate of rehospitalization in the PG group than in the PJ group in the study by Figueras et al, but this difference was not statistically significant.
