Pancreatic Duct Occlusion: A Safe and Valid Alternative to Duct Anastomosis After Pancreaticoduodenectomy in a Low Volume Centre

Background: Pancreaticoduodenectomy is the only possible choice of treatment for peri-ampoullar neoplasms. Morbidity in pancreatic surgery is mainly related to the development of a postoperative pancreatic stula (POPF). According to International Study Group on Pancreatic Fistula it is possible to grade POPF based on clinical variables. Three main different surgical strategies have been proposed to deal with the pancreatic stump following pancreaticoduodenectomy: pancreatojejunostomy, pancreatogastrostomy and pancreatic duct occlusion, but none of them has been clearly demonstrated to be superior to the others. The aim of our study is to evaluate the feasibility of duct occlusion and its correlations with postoperative pancreatic stula, “brittle diabetes” and overall survival in a low volume centre. We decided to review our previous experience in the light of the recent Covid pandemic where, in our country, it has been forced in many regions to displace treatment of oncological patients in low volume hospitals with limited experience Methods: We retrospectively reviewed 56 consecutive patients, from a prospective maintained database, who underwent Whipple’s procedure from January 2007 to December 2014 in a tertiary Hepatobiliary Surgery and Liver Transplant Unit with a low volume of pancreatic resections. The mean follow-up was 24.5 months. Results: The overall incidence of postoperative pancreatic stula was 66.6%: 15 patients had a Grade A (31.25%), 13 a Grade B stula (27.03%), and 4 (8.3%) suffered from a life-threatening Grade C stula. At the last follow-up, 24 of the 28 patients who were alive (85.6%) habitually used substitutive pancreatic enzyme. Conclusion: Duct occlusion can be a safe alternative to pancreatic anastomosis especially in low volume centres and for those patients (age >75 years, obese, hard pancreatic texture, small pancreatic duct) at higher risk of clinically relevant POPF.

pancreatogastrostomy and pancreatic duct occlusion, but none of them has been clearly demonstrated to be superior to the others. The aim of our study is to evaluate the feasibility of duct occlusion and its correlations with postoperative pancreatic stula, "brittle diabetes" and overall survival in a low volume centre. We decided to review our previous experience in the light of the recent Covid pandemic where, in our country, it has been forced in many regions to displace treatment of oncological patients in low Conclusion: Duct occlusion can be a safe alternative to pancreatic anastomosis especially in low volume centres and for those patients (age >75 years, obese, hard pancreatic texture, small pancreatic duct) at higher risk of clinically relevant POPF.
Trial registration: 'retrospectively registered' Background Surgical resection is the only possible choice of treatment in several pancreatic disorders including malignancies, adenomas, traumas and severe acute and/or chronic pancreatitis [1]. Radical resection is the single most important factor in determining outcomes in patients with pancreatic adenocarcinoma [1][2][3].
Morbidity in pancreatic surgery is mainly related to the development of a postoperative pancreatic stula (POPF). According to International Study Group on Pancreatic Fistula (ISGPF) [8] it is possible to grade POPF based on clinical variables. "A grade" stulas, as called a "biochemical leak" (BL) in update classi cation, does not need any treatment (currently it is not considered a true pancreatic stula) and implies no clinical impact. "B grade" stulas can be managed with medications and only prolong length of hospital stay in association with a clinically relevant condition. "C grade" stulas need operative treatment and might be life-threatening [7]. In high volume centres for pancreatic surgery the overall POPF incidence is around 20% [7][8][9].
Intra-abdominal abscesses, delayed gastric emptying, post-pancreatectomy haemorrhage and sepsis represent additional sources of morbidity. In most cases, however, they occur in association or as a consequence of POPF [10,11].
The incidence of postoperative complications has a signi cant impact on the length of hospital stay, costs, quality of life and chance to start chemotherapy [12][13].
Several different surgical and pharmacological approaches have been proposed to avoid POPF which might be different depending on the experience and preferences at each centre [14].
Three main different surgical strategies have been proposed to deal with the pancreatic stump following PD: pancreatojejunostomy (PJ), pancreatogastrostomy (PG) and pancreatic duct occlusion (DO), but none of them has been clearly demonstrated to be superior to the others [15].
In spite of such detailed reporting of morbidity and mortality following PD, it is still not clear whether is surgeon's experience or hospital volume to rescue patients when a complication occurs [15 ]. If PJ is the procedure of choice in medium/high volume centres, DO could be proposed as a safer alternative in medium/low volume centres, in order to reduce the risk of major postoperative complications.
We decided to review our previous experience in the light of the recent Covid pandemic where, in our country, it has been forced in many regions to displace treatment of oncological patients in low volume hospitals with limited experience [16]. The encouraging results of DO in terms of overall survival, POPF and "brittle diabetes" are here presented.

Study Design
We retrospectively reviewed 56 consecutive patients who underwent Whipple's procedure from January 2007 to December 2014 in a tertiary Hepatobiliary Surgery and Liver Transplant Unit with a low volume of pancreatic resections.
All data were obtained from a prospective maintained database and analysed retrospectively. For this type of study due to its retrospective nature a formal ethical review was exempt and a formal consent is not required.
Eight patients were lost at follow up so the analysis on morbidity was conducted on the 48 patients available with a mean follow-up of 25.4 months.
In all cases DO was performed with Cyanoacrylate glue injection.
We recorded data about medical history, Body Mass Index (BMI), American Society of Anaesthesiologists' (ASA) score, preoperative CA19.9, survival, mean operative time, incidence of POPF, incidence of sepsis, incidence of post-operative haemorrhage, re-laparotomy rate, hospital stay, incidence of preoperative and postoperative diabetes, 30-day and 90-day postoperative mortality, oncological recurrence and pancreatic exocrine function.
Pancreatic exocrine function was evaluated by personal or telephonic interviews assessing any substitutive pancreatic enzyme therapy (yes/no) related to steatorrhea/diarrhea since surgery.

Pre-operative work up
Our preoperative work up consisted of total body CT and/or MRI scan for oncological staging and for the exact determination of tumor size and resectabilty. If total bilirubin was higher than 20 mg/dl a biliary drainage was placed via ERCP in patients whose surgery was not scheduled within two weeks.

Surgical technique
We performed a Whipple procedure with an open approach. : gastrectomy was performed using GIA 90 without pylorus preservation.
After pancreatic resection, we performed DO of the Wirsung duct with Cyanoacrylate glue independently from the stump characteristics. In detail, the pancreatic stump was closed with 3/0 polypropylene stiches during glue polymerization while the catheter inserted in the main pancreatic duct for glue injection was simultaneously removed to obtain a complete duct closure.
We nally performed biliary reconstruction with a Roux-en-Y anastomosis.
Two abdominal drainages were placed (one close to the pancreatic remnant and one in the pelvis).

Postoperative care
All patients stayed at least one day in the intensive care unit (range:1-3 days) and then returned to the ward. Amylase and lipase were routinely monitored either in serum and in the drainage starting from postoperative day 3. POPF was de ned according to ISGPS [8,15].
A cephalosporin + metronidazole regimen was always used. Octreotide 0,1 ml was administered subcutaneously three times a day. In absence of POPF patients were allowed oral intake on postoperative day 5.

Statistical Analysis
Descriptive statistics were collected and reported as whole number (percentage) and mean or median (range).
The Fisher exact test was used to compare categorical data. Differences were considered to be statistically signi cant with P values <0.05.

Results
Baseline characteristics of patients are depicted in Table 1.  No patients underwent a pylorus-preserving procedure nor a vascular resection.
Postoperative results are depicted in Table 2. Mean hospital length of stay was 29 (12-56) days with an overall POPF mortality rate of 8.3%.
Pancreatic duct diameter was < 3 mm in the 70% of cases, our percentage of small wirsung diameter is the highest compared to other authors as depicted in the Table 2.
Our incidence of reoperation was 19.6% (Table 4), it was linked to Grade C POPF in 5 patients (45.4%) The incidence of POPF according to pancreatic texture is depicted in picture 1.
Of the four patients with a Grade C stula, two died, accounting for a mortality rate as high as 50%, whereas postoperative mortality for patients with Grade A and Grade B stula was 7.1% (2/28) (Table 3).  Overall survival of the 48 remaining patients at the last follow-up as above was 58.3% (Table 3).

Discussion
PD represents the ideal treatment for periampullary lesions. Complications related to pancreatic duct reconstruction are still the leading cause of morbidity and mortality.
Mortality can be used as an indicator of quality of surgery for PD, but literature lacks an agreed description of the standard of care for such a complex procedure [23].
Surgical outcomes after PD are better in centres performing more than 50 resections per year with a reported overall mortality of less than 5%, compared with a mortality rate of 12.4% in low-volume centres [25,28]. Centres can be considered for pancreatic surgery accreditation if they meet the requirement of 50 pancreatic procedures (including PD) over 3 years with a mortality rate lower than 50% [28].
In a recent study of Krautz et al., the mortality rates reported in Germany ranged from 6.5% in very highvolume hospital to 11.5% in very low-volume ones [30]. Considering only high-volume centres, mortality rates can be below 2% [31].
Similarly, the overall mortality rate in more than 1500 PD performed in Italy was reported to be as high as 8.1% [25]. The authors classi ed hospitals according to volume in low-volume, (< 5 PD/year), mediumvolume (6-13 PD/year), high-volume (14-51 PD/year) and very high-volume (> 90 PD/year) centres, and found that post-operative mortality rate decreased progressively from 12% (low-volume hospitals) to 2.6% (very high-volume hospitals) [25].
Our results show an overall pancreatic surgery-related mortality, as high as 8.3%, which is lower compared to the observed mortality for low-volume centres [25].
However, in the experience of a high-volume centre, postoperative mortality after PJ seemed to be higher than after DO (6.8% vs 2.4%) [32].
Pedrazzoli in a large systematic review on Pancreaticoduodenectomy and pancreatic stula analysed 162 articles involving 54,232 patients. The review shows 4813 Grade A (8.9%), 4830 Grade B (8.9%), and 1872 Grade C (3.5%) POPFs with a mean overall stula rate of 21.3%. A huge variability of Grades A and B POPFs varied from less than 2% to more than 20% with a minimum of 0% and a maximum of 42.5% for Grade A and a minimum of 0.7% and a maximum of 33.3% for Grade B POPF. Grade C POPFs arises from 1% to more than 9% with a maximum of 13.6% [21].
It has been suggested that avoiding an anastomosis of the pancreatic duct by means of duct occlusion could minimize anastomosis-related morbidity, especially in low-volume centres [17,18,32,3]. The aim was to obtain a "pure" pancreatic stula with no activation by bile and/or enteric juice, thereby reducing the risk of life-threating complications.
Di Carlo et al showed that DO procedure was feasible and less time-consuming than PJ, although it could be associated with higher stula rates. However, POPF could not be clinically relevant probably due to the absence of a pancreatic enzymes activation [33].
In our experience the overall incidence of POPF was 66.6%. This observation is consistent with the experience of Tersigni et al, who observed a higher rate of POPF after DO (45.4%) compared to end-to-end PJ anastomosis (15.6%) and to end-to-side PJ anastomosis (11.3%), with a similar incidence of Grade C stula in all the groups (3.1% after end-to-end PJ anastomosis, 2.3 after end-to-side anastomosis and 3.0% after DO) [32]. In our hands only 4 patients (8.3%) had a life-threating POPF. In a recent study comparing 54 patients operated on in a high-volume centre with 44 patients operated on in a low-volume centre over ve years, there were no statistical differences in the incidence of POPF between the two groups (30% vs 27%, P = 0.826) [18]. All patients in this study had a PJ after PD. Interestingly, the rates of Grade C stula were 25% in the high-volume centres and 17% in the low-volume centre. These gures are slightly higher than those observed in our centre with DO. Others have reported that DO has higher postoperative morbidity and mortality, even if not statically signi cant [17].
Consistent with other reports, in our patients a soft pancreatic texture was associated with a signi cantly higher incidence of POPF (overall 27.1% of POPF with soft pancreas vs. 6.25% of POPF with brotic pancreas, P = 0.0068).
In a recent prospective randomised control study [21] compared POPF following PO in high risk patients for pancreatic stula vs PJ after PD for low risk patients for pancreatic stula, mortality after PO was 5.9% and 2.0% after PJ anastomosis, in our serie 90-day mortality related to signi cant POPF was (3/48) 6%, so mortality might be considered superimposable with other authors who performed DO (Table4) .
He et al. analysed RCTs and OCSs, where were related different treatment of pancreatic stump and maojor outcomes after PD or pylorus-preserving PD for malignant or benign pancreatic tumor, chronic pancreatitis, or extra-pancreatic tumors (periampullary, biliary or duodenal) The objective of the metanalysis was a comparison between PJ and PG using quantitative data on PF and overall complications. PD without anastomosis, or duodenum-preserving pancreatectomy was excluded. ( He et. al) [24] We shall underline metanalysis by He et al. reported a lower mortality index performing PG and PJ, but these data were published by high volume and referral centres for pancreatic surgery [28], the same paper reported data by Duffas et al. showing in their experience an incidence of death after PG 10 (12%) and PJ 7 (10%). A summary of these ndings is depicted in Table 5.
Our incidence of reoperation was quite high 19.6% (Table 4), it was linked to Grade C POPF in 5 patients (45.4% of C grade stula patients) our incidence is similar to what other authors reported in literature either after DO either after PA [20-21-23-27], but as depicted in the Table 5, high volume referral centre showed lower rate of reoperations.
In our opinion, in patients with a higher risk for POPF (soft pancreas, dilated pancreatic duct) DO can be a safer option, ideally suitable in low-volume centres.
Four of our patients (8.3%) had postoperative haemorrhage, and all of them needed return to operative room. Interestingly, in only two patients (50%) haemorrhage was a consequence of POPF (all grade A). In the other two cases the bleeding originated from a small vessel from the portal vein and the gastroepiploic artery. The overall incidence of POPF-related bleeding was 4.2%, which is in line with other experiences [15].
Our length of stay was 28 days, higher than those observed in other experiences [7,18]. Of note, availability of post-discharge opportunities, nancial problems, low human resources and patients wish could affect this gure.
More than 80% of patients needed pancreatic enzymes supplementation due to postoperative pancreatic insu ciency. This facet is consistent with others [15][16][17] however, Tran et al reported that the need for enzyme supplementation one year after surgery was not related to the type of reconstruction [17] Probably, pancreatic exocrine insu ciency is more related to the pancreatic atrophy/ brosis and preoperative texture than to DO or PJ [15][16][17].
In our series, 9% of patients developed brittle diabetes, with only 13 patients (27.1%) developing new onset diabetes. This might con rm that DO has higher risk of new onset diabetes, even if only few patients suffer from an uncontrolled diabetes [15][16][17].
According to Tran et al., the incidence of endocrine insu ciency is signi cantly higher after DO compared with PJ at 3-and 12-month follow-up after surgery (P = 0.001 for both) [17].
It is clear that the outcome of complex surgical procedures may not only rely on technical aspects of surgery, but it is also affected by resource availability. However, some technical aspects can be modi ed and reduce the risk of life-threating postoperative complications even in low/medium volume centres.
PD can be safely performed in low-volume centres if amenities and processes typical of high-volume centres can be replicated in specialized units [34][35].
The ideal concept of reserving pancreatic surgery only to high specialized centres is probably utopian. Geographical limitations, elevated costs for the patients and their relatives, political issues, different regional health-care systems, and the opposition by medical and surgical staff determine the need to perform this surgery even in academic or tertiary referral hospitals with a limited experience in pancreatic surgery [19] but with all the amenities required for very complex surgery.
According to Diaz et al. although most patients who need PD bypass the nearest providing hospital to seek care at a higher-volume hospital, nearly 25% of the patients still underwent PD at a low-volume centre [36].
So, considering criteria published in literature [23][24][25][26], pancreatic surgery should be centralized, this implies unavoidably an increase of interregional mobility and related health care costs, especially for patients from region of southern Italy.
During Covid-19 pandemia, as we know from the survey written by Aldrighetti et al. on HPB surgery in Italy [16], 72.8% of HPB centers showed a reduction of routine elective operations ≥ 50%, if we combine effects of centralization to the effects of Covid-19 pandemia we understand how di cult would be for patients to undergo pancreatic surgery in a quite fast, safe and effective way. In this situation we decided to analyze our outcomes from a low volume center for pancreatic surgery to overcome the impossibility to send patients to pancreatic surgery referral centers, considering their overload, ensuring to patients a high-quality service at the same time. Our approach led us to guarantee effective treatment and safety procedures during the critical pandemic period.
Probably, a surgical alternative such as DO during the phase of PD at higher risk of complications, i.e. the pancreatic anastomosis, can reduce the rates of subsequent morbidity and mortality with similar oncological results.

Conclusions
Despite of we understand that our study is a retrospective, single-centre analysis, we considered consecutive patients who underwent PD and were registered in a prospectively maintained database. We can consider our centre as low volume due to the number of PD per year, but we can be supported by high-volume centre facilities, including a) being a referral centre for hepatobiliary surgery, liver transplantation, advanced colorectal surgery, b) having a dedicated intensive care unit, and c) having interventional radiology and endoscopy available 24 h.
A comparison of DO with other types of pancreatic duct reconstructions should be advisable in order to draw de nitive conclusions, ideally by means of an adequately designed randomized controlled trial (RCT). Given the lack of such high quality evidence, our study has some merits. Our aim was to give an overview of current clinical practice in a "real life" setting, considering that not always a RCT can be performed with a su cient power and that observational studies can be properly used with prospectively maintained series to assess feasibility of a surgical procedures.
In conclusion, DO can be a safe alternative to pancreatic anastomosis especially in low/medium volume centres and for those patients (age > 75 years, obese, hard pancreatic texture, small pancreatic duct) at higher risk of clinically relevant POPF. Declarations -Ethics approval and consent to participate: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
The research related to human use has been complied with all the relevant national regulations, institutional policies and in accordance the tenets of the Helsinki Declaration, and has been approved by the authors' institutional review board or equivalent committee.
Informed consent: Informed consent has been obtained from all individuals included in this study Trial registration: 'retrospectively registered' -Consent to publish Not applicable.
-Availability of data and materials