Research Article - (2023) Volume 24, Issue 5
Received: 13-Oct-2023 Published: 09-Nov-2023, DOI: 10.35841/1590-8577-24.5.825
Context: Postpancreatectomy hemorrhage and postoperative pancreatic fistula are major causes of morbidity and mortality after pancreaticoduodenectomy. This single center prospective cohort study compares postpancreatectomy hemorraghe after pancreaticogastrostomy and pancreaticojejunostomy. Our hypothesis is that postpancreatectomy hemorraghe after pancreaticogastrostomy is more prevalent and severe. Design: Between January 2013 and March 2021, data on all patients undergoing pancreaticoduodenectomy were collected prospectively. In 2013, pancreaticogastrostomy was implemented as the reconstruction technique for all patients. However, due to a high incidence of severe postpancreatectomy hemorraghes the reconstruction technique was changed to pancreaticojejunostomy in 2016. The anastomosis remained consistent between the two methods, utilizing interrupted sutures with a “dunking” technique. Results: The risk for postpancreatectomy hemorraghe increases by 5.6 times when postoperative pancreatic fistula occurs (OR=5.60 95% CI [2.21-15.08], p<0.001). After correcting for postoperative pancreatic fistula, the odds on a grade C postpancreatectomy hemorraghe is 6.91 (95% CI [2.96-17.67], p<0.001) times higher after pancreaticogastrostomy, compared to pancreaticojejunostomy. Resulting in more relaparotomies after pancreaticogastrostomy (28.6%), compared to 8.4% after pancreaticojejunostomy ( p<0.001). Conclusions: Pancreaticogastrostomy has been proposed as a safer reconstruction technique after pancreaticoduodenectomy with less postoperative pancreatic fistula. However, no significant difference in postoperative pancreatic fistula was observed in this study. The severity and incidence of grade C postpancreatectomy hemorraghe were significantly higher after pancreaticogastrostomy compared to pancreaticojejunostomy, resulting in a higher number of relaparotomies.
What is the best reconstruction technique after Pancreaticoduodenectomy, Pancreaticogastrostomy (PG) or Pancreaticojejunostomy (PJ)? This has been the subject of ongoing debate between pancreatic surgeons, with no clear answer despite extensive research. According to the current Evidence Map of Pancreatic Surgery by Probst et al., a living systematic review by the ISGPS (International Study Group of Pancreatic Surgery) comparing PG and PJ, there is no difference in mortality, overall complications, bile leak, delayed gastric emptying, length of hospital stay or operation time. However, the only notable difference is that PG has a lower risk of developing Postoperative Pancreatic Fistula (POPF) (low-grade evidence). On the other hand, the risk for Postpancreatectomy Hemorrhage (PPH) appears to be lower after PJ (moderate-grade evidence) [1].
In our department, PJ was initially the preferred method of reconstruction. This was changed to PG during the Belgian RCT by Baki et al. [2] in 2013. However, due to a high rate of severe PPH, resulting in unstable, critically ill patients that had to be rushed to the operating theatre, it was decided to change the technique back to PJ in 2016.
Data collection
Between January 2013 and March 2021, data on all patients undergoing pancreaticoduodenectomy were collected in a prospective database. Postoperative complications were evaluated and graded using the International Study Group of Pancreatic Surgery definition of 2017 (Table 1) [3]. Additionally, the TNM classification of patients operated on before 2016 was modified to align with the 8th Edition TNM Classification of Malignant tumours [4]. Serum bilirubin level was measured within 48 hours prior to surgery.
Ethics
Informed consent and approval of the local Ethical Committee were acquired (Belgian registration number: BE300201318590). This study was conducted according to the ethical principles stated in the "Declaration of Helsinki" and "Good Clinical Practice."
Standardized surgical technique and perioperative care
A standardized pancreaticoduodenectomy was performed in all patients. All visible vessels in the surface of the pancreatic remnant were sutured using prolene 4/0 or 5/0. The anastomosis was made using a “dunking” technique, utilizing interrupted PDS 4/0 sutures to the small bowel loop for the PJ (Child reconstruction) and to the rear side of the stomach for the PG. In all cases, prophylactic continuous IV Somatostatin® 6mg daily was started immediately after pancreatic neck transection and continued based on the drain output and lipase analysis in the following days. During the surgical procedure, the surgeon utilized a probe to measure the Main Pancreatic Duct (MPD). A pancreatic duct that measured over 3mm was considered widened. The primary surgeon evaluated the consistency of the pancreatic remnant during the resection and categorized it as "soft," "hard," or "inflamed."
After surgery, anticoagulation treatment was separated into three categories. The first, referred to as "prophylactic," involved a once-daily dosage of 4000IE (40mg) of Enoxaparin®. The second, known as "intermediary," required a twice-daily dosage of 4000IE Enoxaparin®. Finally, the "therapeutic" group received a twice-daily dosage of 1mg/kg Enoxaparin® or unfractionated heparin, with a target aPTT between 60 and 80s (Table 2).
Statistical analysis
Data were analyzed using R version 4.1.2 (R Core Team 2021, R Foundation for Statistical Computing, Vienna, Austria). The data are expressed as means (standard deviation) for normally distributed continuous variables and medians [IQR (interquartile range)] for non-normally distributed continuous variables. Categorical data are expressed as numbers (%). The normal distribution of continuous variables is assessed graphically using QQ plots. The independent samples t-test for normally distributed variables and the Mann-Whitney U test for non-normally distributed variables were used to compare means and medians for outcome parameters. The Chi-Square or Fisher's exact test (in case more than 20% of the expected values were below 5) was used to compare categorical data. A logistic regression model was fitted with PPH (any grade versus no) as the outcome and type of anastomosis as the independent variable. Possible confounders were added one by one to the model (each time considering a model with two independent variables, type of anastomosis and one other variable). A similar model was considered for the outcome of grade C PPH versus no PPH or grade A/B PPH. A p-value < 0.05 is considered statistically significant (Table 3).
A total of 274 patients underwent pancreatic surgery, with 84 receiving a PG and 190 receiving a PJ. Analysis of patient characteristics in both groups showed no significant differences in age, BMI, consistency of the pancreatic remnant, serum bilirubin, postoperative anticoagulation regimen, and underlying pathology. However, a slightly higher proportion of patients in the PJ group had a widened MPD (60.0% versus 45.3%), and preoperative biliary drainage was performed more frequently in the PG group (44.6% versus 32.9% in PJ group).
Clinically relevant pancreatic fistula occurred in 7.1% of PG patients and 7.9% of PJ patients, with no significant difference between the two groups. The incidence of postpancreatectomy hemorrhage grades A and B was similar in both groups, but grade C PPH was more frequent in the PG group (23.8% compared to 5.3% in PJ group) and required reoperation more often (28.6% compared to 8.4% in PJ group). The mortality rate for POPF- and PPH-related complications was low and not significantly different between the two groups.
Logistic regression analysis showed that the odds of PPH were more than five times higher when POPF occurred, and the type of anastomosis was a significant factor in all models. After adjusting for POPF, the odds of any grade of PPH were 2.51 times higher in the PG group (p=0.002) and the odds of grade C PPH were 6.91 times higher (p<0.001) compared to the PJ group.
After undergoing Pancreaticoduodenectomy, POPF and PPH are common causes of postoperative complications, which could lead to significant morbidity and mortality. In the early nineties, PG was proposed as a better reconstructive technique with a lower risk of POPF, compared to PJ, by some authors [5, 6, 7]. Since then, numerous Randomized Controlled Trials (RCTs) have been conducted, but the results have been contradictory. While some RCTs have reported similar POPF rates for both PG and PJ [8, 9, 10, 11], others have shown a lower incidence of postoperative complications after PG, including postoperative collections [9] and a significantly lower POPF rate of 4% compared to 18% after PJ (p<0.010) [12]. However, none of these RCTs have reported on PPH.
In a 2013 Belgian multicenter randomized controlled trial, 329 patients were evaluated and the results showed a significantly lower rate of grade B or C POPF after PG compared to my (PJ) (8% versus 19.8% after PJ, p<0.010). The study also found no significant difference in PPH between the two techniques (13% after PG and 10% after PJ, p=0.490). The authors of the study propose several mechanisms that may contribute to the lower incidence of POPF after PG, such as less activation of exocrine pancreatic secretions in the acidic gastric environment and better vascularization of the gastric wall compared to the jejunal wall. They also suggest that PG may carry a lower risk of bleeding because it is anatomically further from major vessels. Another RCT conducted by (Figure 1) et al. reported similar outcomes, with a lower POPF rate after double-layered PG (15% after PG compared to 34% after duct-to-mucosa PJ, p=0.020) and a trend towards higher PPH after PG (20% compared to 12% after PJ, p=0.230), although this difference was not statistically significant [13]. Wellner et al. reported a significantly higher incidence of PPH after PG (7% versus 2% after PJ, p=0.36) in anan RCT including 116 patients. No significant difference in POPF was found [14].
Figure 1. PPH according to anastomosis type.
The RECOPANC trial included 320 patients and reported no difference in POPF (20% after PG compared to 22% after PJ, p=0.62). However, they found a significant difference in PPH of 21.1% after PG, compared to 11.4% after PJ (p=0.020) [15].
The high occurrence of grade C bleeding and reoperation after PG could be due to the difference in stomach volume compared to the jejunum. In cases where bleeding occurs after PJ, the small intestine fills up faster, which leads to a tamponade effect on the bleeding pancreatic remnant. This results in less severe bleeding with a more stable hemodynamic state, giving enough time for treatments such as embolization or conservative measures such as pausing anticoagulants and administering tranexamic acid. Another explanation could be the stomach's acidic environment, which may erode the hemostatic stitches on the pancreatic surface. Despite routinely giving our patients proton pump inhibitors, this could still be a contributing factor.l be filled more quickly, resulting in a tamponade of the bleeding pancreatic remnant. This results in less severe bleeding with a more hemodynamically stable patient giving sufficient time to go to the angiography suite for embolization or to treat the bleeding conservatively by pausing anticoagulants and giving tranexamic acid. Another explanation may be that the stomach’s acid environment erodes the hemostatic stitches on the pancreatic surface, although we routinely give our patients a proton pump inhibitor.
This study is a single-center prospective cohort study, which is and an important risk factor for bias [Table 1] shows that the two groups are comparable in many aspects, except for the number of patients with an MPD ≥ 3mm and the presence of preoperative biliary drainage (which has a trend towards significance p<0.100). The slightly higher proportion of widened MPD in the PJ group may attribute to the comparable number of POPF after PG and PJ. However, there is no difference in the consistency of the pancreatic remnant in both groups.
Another difference between both groups is the number of patients that underwent preoperative biliary drainage. This number is lower in the PJ group. This is caused by the fact that in recent years, we have attempted to avoid manipulation of the bile ducts by prioritizing urgent surgery for jaundiced patients. The use of both external and internal drainage methods carries a higher risk of inducing pancreatitis and cholangitis [16, 17].
The main strength of this data series, being a singlecenter experience, is that all anastomoses were performed consistently using interrupted PDS 4/0 sutures and a "dunking" invaginating technique.
The odds on any grade of PPH is 2.51 (95% CI [1.41- 4.47], p=0.002) times higher in the PG group, compared to PJ, after correcting for POPF. The odds on a grade C PPH is 6.91 (95% CI [2.96-17.67], p<0.001) times higher with PG after correcting for POPF. More PPH cases required relaparotomy after PG (28.6%), compared to 8.4% after PJ (p<0.001).
Irrespective of the reconstruction technique, the odds of any grade PPH is more than five times higher when POPF occurs (OR=5.60 95% CI [2.21-15.08], p<0.001).
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Indexed at, Google Scholar, Cross Ref
Copyright: This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
