Independent risk factors for clinically relevant postoperative pancreatic fistula after robotic pancreatoduodenectomy: the protective role of the modified Blumgart pancreatojejunostomy in a single surgeon retrospective cohort of 220 cases

Introduction

Pancreatoduodenectomy (PD) is considered one of the most challenging operations in digestive surgery due to a complex anatomy and the need for multiple and technically demanding reconstructions. It remains the standard treatment for benign and malignant lesions of the pancreatic head but carries substantial morbidity largely due to clinically relevant postoperative pancreatic fistula (CR-POPF).

CR-POPF is the most feared complication of the procedure, increasing hospital morbidity and mortality (1). Indeed, CR-POPF markedly increases the incidence of postoperative hemorrhage, infection, prolonged hospital stay, and the need for reoperation or interventional therapy (2).

Over the past decade, robot-assisted PD (RPD) has been increasingly adopted. Several specialized centers have reported their experience, confirming the feasibility and safety of this minimally invasive approach (3-6). Nevertheless, RPD remains a challenging procedure, and its wider implementation might be tempered by concerns regarding its potential complications especially CR-POPF, and regarding the long learning curves reported from some expert centers.

To date, only a limited number of studies have focused on risk factors for CR-POPF after RPDs (7). Moreover, the results of studies evaluating the impact of the learning curve on the occurrence of CR-POPF are scarce and sometimes contradictory (5,8-11). Further evidence is required to clarify these issues and to inform the safe implementation of RPD, particularly with regard to the type of pancreatic anastomosis.

The objective of this study was to identify independent predictors of CR-POPF after RPD in a large single-surgeon cohort, with specific attention to the role of anastomotic technique and the implementation phase of the program. We present this article in accordance with the STROCSS reporting checklist (available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-2025-712/rc).

Methods

Data were retrospectively collected from medical records of all consecutive patients who underwent RPD at the University Hospital of Orleans, France, between October 20, 2015 and December 31, 2024. All procedures were performed by a single senior surgeon (O.S.M.). Preoperative work-up included general health assessment, contrast-enhanced three-phase computed tomography scan, and routine laboratory tests, including tumor markers when indicated. Liver magnetic resonance imaging was performed when liver metastases were suspected. Surgical indication was confirmed in a multidisciplinary tumor board. Preoperative biliary drainage was performed in patients with serum bilirubin levels superior to 250 µmol/L, and nutritional optimization was undertaken when required.

Definitions

Postoperative pancreatic fistula (POPF), delayed gastric emptying (DGE), and post-pancreatectomy hemorrhage (PPH) were defined according to international study groups criteria (12,13). POPF was graded according to the 2016 International Study Group for Pancreatic Surgery (ISGPS) update (14). CR-POPF was defined as grade B or C fistula.

The updated alternative pancreatic fistula risk score (ua-FRS) was calculated based on sex, body mass index (BMI), pancreatic texture (soft vs. hard), and main pancreatic duct diameter (15).

Postoperative complications were classified using the Clavien-Dindo classification (16). Resection margins were reported according to international criteria, with R0 resection defined as ≥1 mm clearance (17,18).

Surgical technique

Procedures were conducted using the da Vinci robot platform, following a standardized stepwise technique previously described (19,20). Three types of pancreatic anastomosis were used: (I) pancreatogastrostomy; (II) conventional pancreatojejunostomy; and (III) modified Blumgart pancreatojejunostomy. The modified Blumgart technique consisted of a duct-to-mucosa anastomosis with separate 5.0 absorbable monofilament sutures over a short stent, combined with transpancreatic 3.0 absorbable monofilament mattress sutures anchoring the pancreatic stump to the jejunum. A schematic illustration of the modified Blumgart pancreatojejunostomy as performed in our institution is provided in Figure 1.

Figure 1 The modified Blumgart technique. (A) Posterior layer of the modified Blumgart pancreaticojejunostomy, showing three transfixing 3-0 absorbable monofilament sutures placed through the posterior pancreatic parenchyma and tied anteriorly. (B) Posterior layer of the Wirsung-jejunostomy, illustrating placement of the stent within the pancreatic duct and interrupted 5-0 absorbable monofilament sutures securing the duct to the jejunal mucosa. (C) Anterior layer of the Wirsung-jejunostomy anastomosis showing interrupted 5-0 absorbable monofilament duct-to-mucosa sutures. (D) Anterior layer of the modified Blumgart technique, with the previously placed transfixing 3-0 absorbable monofilament sutures brought anteriorly and tied over the pancreatic capsule.

Postoperative management

Patients were managed according to a standardized postoperative protocol applied uniformly to all cases. The nasogastric tube was removed at the end of the procedure, and all patients were admitted to intensive care unit for the first 72 hours. Somatostatin analogues were not administered routinely but were used selectively in cases of persistent POPF beyond postoperative day 5. Drain and serum amylase and lipase were measured on postoperative days 1, 3, and 5. Drains were removed if there was no evidence of biochemical leak and the patient was clinically stable. Oral feeding was initiated on postoperative day 2 and advanced progressively according to tolerance.

Statistical analysis

Continuous variables were expressed as mean and standard deviation (SD), or median and interquartile range (IQR) depending on distribution, and categorical variables as counts and percentages. Group comparisons used Student’s t-test or Mann-Whitney U test for continuous variables, and Pearson’s χ² test or Fisher’s exact test for categorical variables. Linear regression and correlation were applied to assess associations between continuous variables.

Univariate logistic regression identified potential predictors of CR-POPF.

The ua-FRS score (15), which incorporates sex, BMI, main pancreatic duct size, and pancreatic texture, was included in the multivariable logistic regression model together with variables showing a P<0.05 in univariate analysis that were not components of the score.

Because surgical technique and year of surgery were highly correlated, they were not included in the same model.

The learning curve was assessed using cumulative sum (CUSUM) analysis, with associations between case sequence and CR-POPF analyzed using generalized linear modeling, operative time and estimated blood loss analyzed using Pearson correlation.

The proficiency learning curve was evaluated using the risk-adjusted CUSUM (RA-CUSUM) analysis for major complications (Clavien-Dindo ≥ III). The mastery learning curve was assessed using the RA-CUSUM analysis for textbook outcome, defined according to Müller et al. (11) as: hospital stay shorter within the 75^th percentile; absence of mortality; no complication requiring intensive care unit admission; and no reoperation.

All analyses were two-sided with P<0.05 considered significant. Statistical analyses were performed using R software (R Foundation for Statistical Computing, Vienna, Austria, version 4.1.0).

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study protocol was approved by the Ethics Committee of the University Hospital of Orléans (approval No. CHU-ORLÉANS-2023-0084) and individual patient consent was waived for this retrospective analysis.

Results

A total number of 220 patients underwent RPD during the study period. The overall rate of CR-POPF was 17% (n=36). Patients were divided into two groups according to the occurrence of CR-POPF: the fistula (F) group and the no-fistula (NF) group. Patient demographics and peri-operative variables were then compared between groups (Table 1).

The F group included more male patients (78% vs. 59%, P=0.042), had higher ua-FRS score (median 46 vs. 30, P=0.009), and less frequent preoperative biliary drainage (47% vs. 66%, P=0.03). A smaller pancreatic duct diameter was also more common in the F group (<3 mm in 58% vs. 40%, P=0.043). Operative time was similar between groups (median 343 vs. 320 min, P=0.70), and vascular resections were less frequent in the F group, although not statistically significant (0% vs. 18%, P=0.13). The use of a pancreatic duct stent was comparable between groups (71% vs. 69%, P=0.80). There was a trend towards a higher frequency of soft pancreatic texture in the F group (86% vs. 67%, P=0.06), though assessment was limited by 13.6% missing data.

The distribution of anastomotic techniques differed significantly between groups: pancreatojejunostomy was more frequent in the F group (44% vs. 20%), whereas the modified Blumgart technique was less common (47% vs. 71%) (P=0.002).

Univariate analysis

Factors associated with CR-POPF included sex, higher ua-FRS score, absence of preoperative drainage, pancreatic duct<3mm and pancreatojejunostomy.

Multivariate analysis

Independent risk factors for CR-POPF were ua-FRS and pancreatojejunostomy (Table 2). Among these, pancreatojejunostomy was the strongest predictor [odds ration (OR) 3.06, 95% confidence interval (CI): 1.12-8.40, P=0.03].

Learning curve

For CR-POPF, two distinct phases were observed. Before the 68^th case, multiple anastomotic techniques were used, and CR-POPF rates were higher and more heterogeneous. After the 68^th case, coinciding with the near-exclusive adoption of the modified Blumgart technique, CR-POPF rates progressively decreased (P=0.04) (Figure 2).

Figure 2 CUSUM curve of clinically relevant POPF showing a significant learning effect, with an inflection point around the 68^th case. CR-POPF, clinically relevant postoperative pancreatic fistula; CUSUM, cumulative sum.

CUSUM analysis did not show a significant learning effect for intraoperative estimated blood loss (P=0.21) and demonstrated only a non-significant trend for operative time (P=0.56) (Figure 3A,3B).

Figure 3 Multidimentional learning curve analysis of robotic pancreatoduodenectomy. (A) Feasibility learning curve based on operative time. (B) Feasibility learning curve based on estimated blood loss. (C) Proficiency learning curve using the risk-adjusted CUSUM for major complications (Clavien-Dindo ≥ III). (D) Mastery learning curve using the risk-adjusted CUSUM for textbook outcome. CUSUM, cumulative sum; EBL, estimated blood loss; RA-CUSUM, risk-adjusted CUSUM.

RA-CUSUM analysis for major complications (Clavien–Dindo ≥ III) identified the proficiency inflection point at the 68^th case (Figure 3C).

The Ra-CUSUM analysis for textbook outcome identified the mastery phase at the 67^th case (Figure 3D).

Distribution of anastomosis types

The distribution of anastomosis types over time is shown in Figure 4. The first phase demonstrated marked heterogeneity in the anastomosis types performed, whereas the second phase was characterized by systematic use of the modified Blumgart pancreatojejunostomy.

Figure 4 Distribution of pancreatic anastomosis types over time, illustrating the progressive transition towards standardization of the modified Blumgart anastomosis between the 60^th and 80^th cases.

Discussion

PD remains a technically demanding procedure, and CR-POPF continues to be its most serious complication. Known risk factors include elevated BMI, male sex, a small pancreatic duct, and soft pancreatic texture. However, the optimal reconstruction technique after PD is still debated, and no consensus exists regarding the type of pancreatic anastomosis to be performed.

With the development of robotic surgery, PD can now be performed through a minimally invasive approach. Several specialized centers have demonstrated the feasibility and safety of RPD, with outcomes comparable to open surgery (3-6). Nevertheless, RPD raises specific questions, including whether known risk factors for CR-POPF in open PD apply to the robotic setting, and whether the type of anastomosis impacts outcomes. Randomized data remain scarce, and available series report conflicting results regarding CR-POPF incidence and the effect of the learning curve (21-23).

In this study, conducted from the very beginning of our RPD program, we identified two independent predictors of CR-POPF: the ua-FRS score and the type of anastomosis. Conventional pancreatojejunostomy was associated with a significantly increased risk of CR-POPF compared with the modified Blumgart technique (OR 3.06, 95% CI: 1.12–8.40). These findings are consistent with recent reports highlighting the protective role of the Blumgart reconstruction in both open and minimally invasive PD (3,5,23). In our cohort, male sex and a main pancreatic duct diameter <3 mm were associated with CR-POPF, whereas BMI and pancreatic texture were not.

To our knowledge, this is the first study to analyze risk factors for CR-POPF after RPD while explicitly accounting for the implementation phase of the program.

These findings must, however, be interpreted in the context of the learning process. Distinguishing the impact of the learning curve from the effect of the anastomotic technique is inherently challenging in RPD. In our cohort, the transition toward near-exclusive adoption of the modified Blumgart anastomosis coincided with the attainment of both the proficiency and mastery phases on RA-CUSUM analysis. Although increasing surgical experience undoubtedly contributed to improved outcomes, the temporal alignment of these milestones suggests that standardization of the anastomosis played an independent and meaningful role.

Our study is the first to demonstrate that modifying the anastomotic technique may have a more direct influence on postoperative outcomes than the learning curve alone. In our experience, this transition allowed us to reach the proficiency level earlier than reported in international benchmarks (10), and accelerated the attainment of the mastery phase, reached around the 65^th case.

In contrast, CUSUM analyses of operative time and blood loss did not show significant inflection points (P=0.56 and P=0.21, respectively). Although operative time showed a non-significant trend, these indicators were less informative than CR-POPF, severe morbidity and textbook outcomes for evaluating program maturation. Unlike previous studies that defined learning-curve phases primarily based on operative time or blood loss (5,11), our findings underscore the predominant role of anastomotic standardization.

The strengths of this study include its large, consecutive cohort and the consistency afforded by a single surgeon’s experience which ensures homogeneity and avoids inter-operator variability. A limitation, however, is that the study reflects the experience of a single surgeon. While this enhances internal validity, external validation remains necessary. Encouraging preliminary results have been observed in a second surgeon who adopted the modified Blumgart technique from case one, but the number of procedures remains insufficient for inclusion in the present study. Other limitations relate to the retrospective design and missing data on pancreatic texture. These missing data reflect the intrinsic subjectivity of assessing pancreatic consistency, a challenge further amplified by the lack of haptic feedback in robotic surgery.

Conclusions

Our findings demonstrate that the modified Blumgart pancreatojejunostomy is independently protective and, once standardized, substantially reduces CR-POPF rates. These results support its systematic adoption to enhance safety and facilitate more efficient learning curve progression in centers implementing RPD.

Acknowledgments

The authors thank Dr. Daniel Pietraz for creating the schematic illustration of the modified Blumgart pancreatojejunostomy presented in Figure 1.

Footnote

Reporting Checklist: The authors have completed the STROCSS reporting checklist. Available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-2025-712/rc

Data Sharing Statement: Available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-2025-712/dss

Peer Review File: Available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-2025-712/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-2025-712/coif). The authors have no conflicts of interest to declare.

Ethical Statement: 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 study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study protocol was approved by the Ethics Committee of the University Hospital of Orléans (approval No. CHU-ORLÉANS-2023-0084) and individual patient consent was waived for this retrospective analysis.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Schmidt CM, Choi J, Powell ES, et al. Pancreatic fistula following pancreaticoduodenectomy: clinical predictors and patient outcomes. HPB Surg 2009;2009:404520. [Crossref] [PubMed]
2. Yeo CJ, Cameron JL, Sohn TA, et al. Six hundred fifty consecutive pancreaticoduodenectomies in the 1990s: pathology, complications, and outcomes. Ann Surg 1997;226:248-57; discussion 257-60. [Crossref] [PubMed]
3. Giulianotti PC, Sbrana F, Bianco FM, et al. Robot-assisted laparoscopic pancreatic surgery: single-surgeon experience. Surg Endosc 2010;24:1646-57. [Crossref] [PubMed]
4. Zureikat AH, Beane JD, Zenati MS, et al. 500 Minimally Invasive Robotic Pancreatoduodenectomies: One Decade of Optimizing Performance. Ann Surg 2021;273:966-72. [Crossref] [PubMed]
5. Shi Y, Wang W, Qiu W, et al. Learning Curve From 450 Cases of Robot-Assisted Pancreaticoduocectomy in a High-Volume Pancreatic Center: Optimization of Operative Procedure and a Retrospective Study. Ann Surg 2021;274:e1277-83. [Crossref] [PubMed]
6. de Graaf N, Zwart MJW, van Hilst J, et al. Early experience with robotic pancreatoduodenectomy versus open pancreatoduodenectomy: nationwide propensity-score-matched analysis. Br J Surg 2024;111:znae043. [Crossref] [PubMed]
7. Polanco PM, Zenati MS, Hogg ME, et al. An analysis of risk factors for pancreatic fistula after robotic pancreaticoduodenectomy: outcomes from a consecutive series of standardized pancreatic reconstructions. Surg Endosc 2016;30:1523-9. [Crossref] [PubMed]
8. Zwart MJW, Nota CLM, de Rooij T, et al. Outcomes of a Multicenter Training Program in Robotic Pancreatoduodenectomy (LAELAPS-3). Ann Surg 2022;276:e886-95. [Crossref] [PubMed]
9. Zhang T, Zhao ZM, Gao YX, et al. The learning curve for a surgeon in robot-assisted laparoscopic pancreaticoduodenectomy: a retrospective study in a high-volume pancreatic center. Surg Endosc 2019;33:2927-33. [Crossref] [PubMed]
10. Jones LR, Zwart MJW, de Graaf N, et al. Learning curve stratified outcomes after robotic pancreatoduodenectomy: International multicenter experience. Surgery 2024;176:1721-9. [Crossref] [PubMed]
11. Müller PC, Kuemmerli C, Cizmic A, et al. Learning Curves in Open, Laparoscopic, and Robotic Pancreatic Surgery: A Systematic Review and Proposal of a Standardization. Ann Surg Open 2022;3:e111. [Crossref] [PubMed]
12. Wente MN, Bassi C, Dervenis C, et al. Delayed gastric emptying (DGE) after pancreatic surgery: a suggested definition by the International Study Group of Pancreatic Surgery (ISGPS). Surgery 2007;142:761-8. [Crossref] [PubMed]
13. Wente MN, Veit JA, Bassi C, et al. Postpancreatectomy hemorrhage (PPH): an International Study Group of Pancreatic Surgery (ISGPS) definition. Surgery 2007;142:20-5. [Crossref] [PubMed]
14. Bassi C, Marchegiani G, Dervenis C, et al. The 2016 update of the International Study Group (ISGPS) definition and grading of postoperative pancreatic fistula: 11 Years After. Surgery 2017;161:584-91. [Crossref] [PubMed]
15. Mungroop TH, Klompmaker S, Wellner UF, et al. Updated Alternative Fistula Risk Score (ua-FRS) to Include Minimally Invasive Pancreatoduodenectomy: Pan-European Validation. Ann Surg 2021;273:334-40. [Crossref] [PubMed]
16. Clavien PA, Barkun J, de Oliveira ML, et al. The Clavien-Dindo classification of surgical complications: five-year experience. Ann Surg 2009;250:187-96. [Crossref] [PubMed]
17. Delpero JR, Bachellier P, Regenet N, et al. Pancreaticoduodenectomy for pancreatic ductal adenocarcinoma: a French multicentre prospective evaluation of resection margins in 150 evaluable specimens. HPB (Oxford) 2014;16:20-33. [Crossref] [PubMed]
18. Delpero JR, Jeune F, Bachellier P, et al. Prognostic Value of Resection Margin Involvement After Pancreaticoduodenectomy for Ductal Adenocarcinoma: Updates From a French Prospective Multicenter Study. Ann Surg 2017;266:787-96. [Crossref] [PubMed]
19. Iben-khayat A, Felli E, Thebault B, et al. Short-term results of robot-assisted pancreatoduodeodenectomy: a retrospective cohort study of 146 patients operated in a high-volume center. HPB 2024;26:1270-9. [Crossref] [PubMed]
20. Thebault B, Quetel R, Peyrafort P, et al. Pancréatectomies mini-invasives robot-assistées. EMC - Techniques chirurgicales - Appareil digestif 2024;41:1-13.
21. Klotz R, Mihaljevic AL, Kulu Y, et al. Robotic versus open partial pancreatoduodenectomy (EUROPA): a randomised controlled stage 2b trial. Lancet Reg Health Eur 2024;39:100864. [Crossref] [PubMed]
22. de Graaf N, Emmen AMLH, Ramera M, et al. Minimally invasive versus open pancreatoduodenectomy for pancreatic and peri-ampullary neoplasm (DIPLOMA-2): study protocol for an international multicenter patient-blinded randomized controlled trial. Trials 2023;24:665. [Crossref] [PubMed]
23. Nickel F, Wise PA, Müller PC, et al. Short-term Outcomes of Robotic Versus Open Pancreatoduodenectomy: Propensity Score-matched Analysis. Ann Surg 2024;279:665-70. [Crossref] [PubMed]