How-I-do-it: meso-caval shunt: a safe emergency option for venous reconstruction in pancreaticoduodenectomy

Abstract

Purpose

In pancreaticoduodenectomy, extensive involvement of the superior mesenteric vein (SMV) often complicates venous reconstruction, especially when artificial grafts are unavailable. In such cases, meso-caval shunt (MCS) emerges as a solution. This study aimed to evaluate the safety and efficacy of permanent MCS as both an emergency procedure and a planned surgical strategy.

Methods

We retrospectively analyzed 8 consecutive patients undergoing pancreaticoduodenectomy with permanent MCS between July 2019 and January 2025 at our institution. Clinicopathological characteristics were identified using electronic medical records, and the same surgeon performed all surgeries.

Results

The study included seven cases of pancreatic cancer and one ampullary carcinoma (6 male, 2 females; age 54–71 years). The mean operative duration was 510 min with 767 ml blood loss. All patients achieved R0 resection with minor complications (Clavien-Dindo grade I/II), no mortality, and preserved portal perfusion. Meso-caval shunt patency was confirmed postoperatively without hepatic dysfunction. Median survival reached 21 months.

Conclusion

Permanent MCS is a reliable elective option for venous reconstruction in pancreaticoduodenectomy, especially when graft-based solutions are not available. Its success in achieving R0 resection while maintaining portal hemodynamics suggests its utility in complex pancreatic resections.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00423-025-03826-1.

Introduction

Pancreatic ductal adenocarcinoma (PDAC) was ranked third among cancer-related deaths in 2023, with a 1.3% annual growth rate [1]. Despite significant advancements in pancreatic cancer treatment over the past decade, including surgical, radiotherapeutic, chemotherapeutic, and immunotherapeutic approaches, the five-year overall survival rate remains dismal at just 12% ^1,2. Although resection is the most effective method of increasing survival, only 15–20% of patients are eligible for surgery at diagnosis, with more than half presenting with distant metastases [2, 3]. According to the National Comprehensive Cancer Network (NCCN) guidelines, PDAC is classified into three stages based on tumor location and proximity to critical vascular structures: resectable, borderline resectable, and locally advanced PDAC [4]. Among these, venous involvement—particularly of the superior mesenteric vein (SMV)—remains a major technical challenge. In pancreaticoduodenectomy (PD), extensive SMV involvement often requires venous reconstruction. However, the use of artificial grafts is frequently impractical due to limited availability or patient-specific constraints. The meso-caval shunt (MCS) offers a viable alternative, particularly in emergencies where conventional in situ vascular reconstruction is unfeasible. This study aims to assess the safety and efficacy of meso-caval shunting in pancreaticoduodenectomy, especially when artificial grafts are unavailable, and to detail our meso-caval shunt surgical approach for PD in cases involving the superior mesenteric vein.

Materials and methods

Study patients

This retrospective single-institution case series study was conducted at China-Japan Friendship Hospital. The study adhered to the principles of the Declaration of Helsinki and received approval from the Ethics Committee of the China-Japan Friendship Hospital (approval number: KY2024-370-01). Written informed consent was obtained from all participants after a detailed explanation of the risks associated with meso-caval shunt during pancreaticoduodenectomy. We identified eight consecutive adult patients who underwent PD with MCS through an electronic medical record review conducted between July 2019 and January 2025. All surgical procedures were performed by the same attending surgeon (Fig. 1). This case series has been reported in line with the Preferred Reporting Of Case Series in Surgery (PROCESS) guideline [5].

Fig. 1.

Flowchart of included patients

Surgical procedure

MCS was employed when conventional venous reconstruction methods were rendered unfeasible by extensive SMV involvement or unavailable artificial grafts. Our technical approach emphasized three critical principles: (1) preservation of portal flow, especially portal-splenic vein confluence patency, (2) tension-free anastomosis, and (3) minimization of clamp time.

Preoperative evaluation included contrast-enhanced computed tomography (CT) with three-dimensional (3D) reconstruction to assess venous involvement, IVC-SMV anatomical relationship, and surgical feasibility. Intraoperatively, abdominal exploration first confirmed the splenic vein (SV) involvement status and evaluated the technical feasibility of reconstructing the SV-portal vein (PV) confluence post-resection, ensuring preserved hepatic perfusion. The SMV resection margin was then meticulously measured to ensure adequate residual length for tension-free anastomosis with the inferior vena cava (IVC) (Fig. 2A). Definitive surgical steps included:

Fig. 2.

A: Representative pictures of patient with meso-caval shunt before the stage of the SMV resection; B: The tumor-invaded SMV has been removed; C: The SV-PV was reconstructed in situ and an end-to-side anastomosis was performed between the SMV and IVC. * PV, portal vein; SMV, superior mesenteric vein; SV, splenic vein; IVC, inferior vena cava;

1. SMV mobilization: Dissection along the pancreatic inferior border to achieve a sufficient free SMV segment for anastomosis (Fig. 2B).

2. Pancreaticoduodenectomy: En bloc resection (including head of the pancreas, duodenum and the involved PV-SMV) using the arterial-first technique [6]with continuous monitoring of SV backflow.

Partial SV resection necessitated immediate in situ SV-PV reconstruction using 6 − 0 polypropylene sutures.

• 3.Meso-caval shunt construction: After partial IVC clamping (30–50% circumference), an elliptical venotomy was created on the IVC anterior wall. End-to-side SMV-IVC anastomosis was completed with 5 − 0 polypropylene running sutures under loupe magnification (Fig. 2C), limiting clamp time to less than 20 min.
• 4.Intestinal viability assessment: The status of the intestinal blood supply guided decisions for right hemicolectomy.
• 5.Digestive reconstruction: Standard pancreaticojejunostomy (end-to-side), hepaticojejunostomy, and antecolic gastrojejunostomy were performed.

Postoperative management

All patients received protocolized care cefoperazone sulbactam (3 g q12h), omeprazole (40 mg q12h), octreotide (600 µg qd × 7 days), and low molecular weight heparin (3000 IU qn). Serial liver function tests were performed on postoperative day 1, 3, and 5. Vascular patency was evaluated via abdominal-enhanced CT two weeks postoperatively. Semiannual clinical surveillance was mandated during the first postoperative year. The date of the last follow-up was January 2025.

Data collection

Baseline demographics, intraoperative characteristics, and outcomes were extracted from the electronic medical records. Specific demographics included age at the time of surgery and gender. R0 resection margin was defined as no tumor cells were identified at any of surgical resection margins. Various laboratory variables were obtained, including preoperative carbohydrate antigen199 (CA19-9), preoperative carbohydrate antigen125 (CA125), postoperative carbohydrate antigen199 (CA19-9), postoperative carbohydrate antigen125 (CA125), postoperative alanine transaminase (ALT) in postoperative day 1/3/5/30 (POD 1/3/5/30), total bilirubin (TBIL) in POD 1/3/5/30, prothrombin time (PT) in POD 1/3/5/30. The Clavien-Dindo classification was used to define postoperative complications [7]. Postoperative pancreatic fistula (POPF) [8]delayed gastric emptying (DGE) [9]and postpancreatectomy hemorrhage (PPH) [10] were consistent with the definitions by the International Study Group for Pancreatic Surgery (ISGPS).

Results

A total of eight patients underwent permanent meso-caval shunt (MCS) placement during pancreaticoduodenectomy between July 2019 and January 2025. Table 1 summarizes perioperative and intraoperative characteristics. The cohort comprised six males and two females aged 54–71 years. Histopathological diagnoses included pancreatic ductal adenocarcinoma (n = 7) and ampullary carcinoma (n = 1). Six borderline resectable pancreatic cancer cases underwent multidisciplinary team (MDT) discussion, which recommended radical resection as primary treatment after declining neoadjuvant therapy. Neoadjuvant regimens were administered to patients with hepatic metastases (n = 1) and ampullary carcinoma (n = 1) before surgery upon meeting resection criteria (Table S1).

Table 1.

Perioperative and intraoperative characteristics of the patients

Case	Age/Gender	Diagnosis	Operation	Invasion of SMV	MCS Indications	OR time	Complication	State (month)
1	64y/M	AC	PD + MCS + RHC	5 cm	grafted unavailable	420 min	None	Dead, 12
2	71y/M	PC + LM	PD + MCS + PH	5 cm	grafted unavailable	600 min	PPH grade A	Dead, 7
3	54y/F	PC	PD + MCS	3.5 cm	grafted unavailable	630 min	None	Dead, 30
4	60y/M	PC	PD + MCS	5 cm	grafted unavailable	510 min	None	Dead, 21
5	58y/M	PC	PD + MCS	2.5 cm	grafted unavailable	430 min	None	Alive, 23
6	70y/M	PC	PD + MCS + RHC	4 cm	grafted unavailable	470 min	None	Dead, 11
7	69/F	PC	PD + MCS	3 cm	Emergency	290 min	None	Alive, 7
8	67/M	PC	PD + MCS + RHC	2 cm	Emergency	760 min	None	Alive, 15

*M Male, F Female, AC ampullary carcinoma, PC pancreatic cancer, LM liver metastases, PD pancreaticoduodenectomy, MCS meso-caval shunt, RHC right hemicolectomy, PH partial hepatectomy, OR time operation time, PPH postpancreatectomy hemorrhage

Mean operation duration was 510 min (range 420–630) with 767 mL (range 200–1200) estimated blood loss. Intraoperative measurements revealed the superior mesenteric vein invasion lengths of 2.5–5.0 cm, with MCS anastomosis completed in less than 30 min. Three patients required right hemicolectomy due to compromised colonic perfusion post-reconstruction. Three out of eight patients (37.5%) developed non-shunt-related complications: two Clavien-Dindo grade I/II events and one grade A postpancreatectomy hemorrhage. No major complications (grade ≥ III) occurred. Postoperative hospitalization spanned 5–30 days across the cohort. (Table S1)

Pathological characteristics are summarized in Table 2. The mean maximum tumor diameter measured 3.73 ± 0.74 cm, with lymph node metastases identified in 5/8 patients (62.5%). All patients achieved a clean margin. Postoperative enhanced CT at 1 month confirmed a patent meso-caval shunt flow (Fig. 3). CA19-9 levels demonstrated > 50% reduction from preoperative baselines in all evaluable cases. Liver function parameters (ALT, TBIL, PT) normalized within 5 postoperative days in most patients, with no significant elevations observed at POD 30 (Fig. 4).

Table 2.

Pathological characteristics

	PD with meso-caval shunt (n = 8)
Diagonosis (%)
Ampulla	1 (12.5)
Pancreatic head carcinoma	8 (87.5)
Maximal diameter of the tumor (mean ± SD, cm)	3.73 ± 0.74
Degree of histological differentiation (%)
Moderately differentiated adenocarcinoma	5 (62.5)
Moderately-poorly differentiated adenocarcinoma	2 (25)
Poorly differentiated adenocarcinoma	1(12.5)
Pathological T stage (%)
T2	4 (50)
T3	4 (50)
Pathological N stage (%)
N0	3 (37.5)
N1-2	5 (62.5)
Harvested lymph nodes (mean ± SD)	26 ± 11.2
Negative margin resection, R0 (%)	8 (100)

Fig. 3.

A: The vessels after surgery in sagittal view of enhanced CT in patient 5; B: Representative enhanced CT with three-dimensional reconstruction shows the vessels after surgery in patient 5; C: Representative enhanced CT with three-dimensional reconstruction shows the vessels after surgery in patient 6; D: The vessels after surgery in sagittal view of enhanced CT in patient 6

Fig. 4.

A: Perioperative alanine transaminase levels of patients with PD and meso-caval shunt; B: Perioperative total bilirubin levels of patients with PD and meso-caval shunt; C: Perioperative prothrombin time levels of patients with PD and meso-caval shunt. *ALT, alanine transaminase; TB, total bilirubin; PT, prothrombin time, PreALT, preoperative alanine transaminase; PreTB, preoperative total bilirubin; PrePT, preoperative prothrombin time; POD, postoperative day;

Five patients died during follow-up, with no 30-day mortality. Median postoperative survival was 21 months. One patient with liver metastatic disease succumbed to cardiovascular disease at two months postoperatively, while another died from pneumonia ten months after surgery. Three deaths resulted from disease progression, while three patients remain recurrence-free under active surveillance.

Discussion

Pancreaticoduodenectomy (PD) has evolved into a safe procedure for pancreatic head tumors and periampullary tumors. However, only one-fifth of PDAC patients present with initially resectable disease due to frequent vascular involvement [3]. Recent advances in systemic therapy (chemotherapy, radiotherapy) [11] and surgical techniques have transformed the management of borderline/locally advanced PDAC from palliative to a multimodal strategies [12]. At high-volume centers, PD with venous resection has become standard practice, with evidence confirming comparable long-term survival outcomes between standard PD and procedures requiring venous resection [13, 14]. The ISGPS classification system further standardizes venous reconstruction into four distinct types [15]. Although interposition grafts essential for extensive venous involvement, traditional options such as falciform ligament grafts(historically employed at out institution [16]) are plagued by unacceptably high rates of late thrombosis. This fundamental limitation compels us to regard in situ repair as just one option among several for managing extensive SMV invasion, and that more technical innovation is required to produce a better therapeutic impact.

Historically, the meso-caval shunt has been employed as a portosystemic shunt for portal hypertension. Christians et al. [17] pioneered temporary meso-caval shunt (tMCS) using internal jugular vein grafts to facilitate PV-SMV dissection in 11 cases, emphasizing intraoperative hemorrhage control. tMCS has also been used after neoadjuvant therapy in borderline resectable pancreatic cancer to minimize intraoperative hemorrhage and prevent intestinal necrosis [18]. As a summary of the management of pancreatic cancer with superior mesenteric vein invasion at Johns Hopkins University, Wolfgang et al. [19] frequently use meso-caval shunt to treat PDAC patients with distal SMV blockage of jejunal and ileal branches.

Our center has innovatively adapted the meso-caval shunt to address extensive PV-SMV tumor infiltration. This graft-free venous reconstruction technique demonstrates particular utility when traditional graft-based approaches are unavailable. Our protocol begins with meticulous preoperative planning utilizing high-resolution CT angiography with three-dimensional reconstruction to precisely evaluate the extent of venous involvement, assess portal-splenic vein confluence patency, and identify optimal anastomotic sites. During the operation, a thorough exploration is performed to confirm that the reflux of PV-SV remains unaffected after the resection and reconstruction of the involved area. After tumor removal, the remaining segments of the mesenteric superior vein, splenic vein, and portal vein are completely liberated to ensure adequate length for anastomosis with the inferior vena cava. As a result, the liver’s blood supply remained intact, and the excision of vessels affected by the tumor allowed for an R0 resection. In addition, systematic assessment of intestinal perfusion is a key step, with concomitant right hemicolectomy performed when tumor involvement necessitates ileocolic artery sacrifice, as required in three of our eight cases. This combined approach does not increase postoperative complication rates while significantly improving surgical exposure for complete tumor resection [20].

Our clinical outcomes demonstrate that MCS represents a feasible alternative for venous reconstruction during PD, particularly when artificial grafts are unavailable. We achieved 100% R0 resection rates with patent shunts confirmed on 30-day follow-up CT angiography. Postoperative liver function remained stable. The median operation time was 540 min (with MCS anastomosis completed in 30 min), and hospital stays ranged from 5 to 16 days following surgery. With no 30-day mortality and only one oncologic death at 60 days (due to cardiovascular diseases), our center achieved no 90-day operation-related mortality. The median survival reached 21 months, surpassing our institutional benchmark of 18 months for resected PDAC.

Whereas most of these cases were temporary meso-caval shunt [21]our center uses them as a means of permanent reconstruction. Unlike approaches requiring internal jugular vein or autologous substitute grafts, our technique reduces graft-related thrombosis [22] and current absence of standardized anticoagulation protocols. In conclusion, our experience demonstrates that permanent meso-caval shunt represents a safe and effective option for malignancies with SMV involvement. Unlike temporary shunts used primarily to facilitate resection [23, 24]our permanent reconstruction technique provides durable venous drainage while avoiding graft-related complications. Critical to success are: preoperative CT angiography to assess SMV involvement and SV-PV patency, preoperative CT angiography to assess SMV involvement and SV-PV patency, and precise preoperative planning of the SMV-IVC anastomotic site. This approach achieves reliable long-term patency without requiring complex anticoagulation management.

While MCS offers a valuable salvage approach, its application requires careful patient selection due to three fundamental limits. First are the anatomic limitations, the procedure proves unsuitable when SMV remnant length is < 2 cm or IVC involvement exists. Second, physiologic contraindications, which significantly increase perioperative risks, especially liver dysfunction (Child-Pugh grade C), Liver cirrhosis, and hypercoagulable states. The final one is technical prerequisites, which call for both institutional expertise and individual surgeon proficiency. These restrictions emphasize the need for meticulous preoperative assessment and coordinated care at high-volume centers.

This study has the following limitations. First, as a retrospective single-center analysis, our study was constrained by incomplete data availability for certain clinically relevant variables (e.g., performance status and comorbidities). Second, the relatively small sample size may introduce selection bias. These methodological constraints highlight the necessity for additional multicenter studies to validate our observations. Thirdly, our study focused on 30-day postoperative outcomes due to retrospective design. While this timeframe captures most acute complications, longer follow-up would better evaluate delayed sequelae such as functional recovery. Future prospective studies should incorporate standardized longer-term assessments to address this limitation. Finally, considering that suitable patients for meso-caval shunt are usually borderline resectable cases, the successful application of this technique may require resources and expertise only available in experienced centers.

Conclusion

The permanent meso-caval shunt technique provides a safe and durable venous reconstruction for pancreaticoduodenectomy, particularly when autologous or prosthetic grafts are unavailable. Our results demonstrate consistent R0 resection rates with preserved portal perfusion and durable patency, supporting its role in managing complex pancreatic resections with SMV involvement. This technique offers a reliable, graft-free option that should be considered in selected patients at specialized centers.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Study conception and design: Zhiying Yang, Ruili WeiAcquisition of data: Ruili Wei, Yue, Qiu, Ruiyi Liu, Hanchun Huang, Wenying ZhouAnalysis and interpretation of data: Ruili Wei, Jia Huang, Li XuDrafting of manuscript: Ruili Wei, Jia Huang, Zhiying YangThe attending surgical team: Zhiying Yang, Yongliang Sun, Jia Huang, Li XuCritical revision of manuscript: Zhiying Yang, Yongliang Sun.

Funding

National High Level Hospital Clinical Research Funding & Elite Medical Professionals Project of China-Japan Friendship Hospital(NO.ZRJY2023-GG04).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.