Vascular resection and reconstruction at pancreatico-duodenectomy: technical issues

Hong Kong, China

Vascular resection and reconstruction at pancreatico-duodenectomy: technical issues

Edward CS Lai

Hong Kong, China

BACKGROUND: With the improvement of perioperative management over the years, pancreatico-duodenectomy has become a safe operation despite its technical complexity. The presence of concomitant visceral artery occlusion unrelated to the underlying malignancy and concomitant major venous infiltration by tumor poses additional hazards to resection which could compromise the postoperative outcome.

DATA SOURCES: A MEDLINE database search was performed to identify relevant articles using the key words "median arcuate ligament syndrome", "superior mesenteric artery", "replaced right hepatic artery", and "portal vein resection". Additional papers and book chapters were identified by a manual search of the references from the key articles.

RESULTS:Computed tomography with 3-dimensional reconstruction of the vascular anatomy provides most key information on the potential vascular problems encountered during surgery. A trial clamping of the gastroduodenal artery provides a simple intraoperative assessment for the presence of any significant visceral arterial occlusion. Depending on the timing of diagnosis, division of the median arcuate ligament, bypass or endovascular stenting should be considered. Portal and superior mesenteric vein resection had been used with increasing frequency and safety. The steps and methods taken to reconstruct the venous continuity vary with individual surgeons, and the anatomical variations encountered. With segmental loss of the portal vein, opinions differs with regard to the preservation of the splenic vein, and when divided, the necessity of restoring its continuity; source of the autologous vein graft when needed and whether the use of synthetic graft is a safe alternative.

CONCLUSIONS: During a pancreatico-duodenectomy, images of computed tomography must be carefully studied to appreciate the changes and variation of vascular anatomy. Adequate preoperative preparation, acute awareness of the probable arterial and venous anatomical variation and the availability of expertise, especially micro-vascular surgery, for vascular reconstruction would help to make the complex pancreatic resection a safer procedure.

(Hepatobiliary Pancreat Dis Int 2012;11:234-242)

pancreatico-duodenectomy; vein resection; celiac axis compression

Introduction

Pancreatico-duodenectomy (PD), or the Whipple procedure, remains a procedure with great complexity. Since the first successful report by Whipple in 1935, the operation took off slowly in the early years because of its prohibitively high mortality rate.[1]Over the years, however, there was already dramatic improvement in the postoperative outcome reported since the 1990s.[2]According to the report on 1423 patients undergoing PD at the Johns Hopkins Hospital between 1970 and 2006, the perioperative mortality rate decreased from 30% in the 1970s to only 1% among the 548 patients operated on since 2000.[3]The improvement of postoperative outcome after PD does not confine to the highly specialized high-volume medical centers. In another review of 2461 Medicare patients with pancreatic cancer of over 64 years old treated between 1991 and 2005 in the United States, a significant reduction of 30-day mortality rate (6% in 1991-1996, to 3% in 2003-2005) was observed despite a higher age and a number of co-morbid factors present among patients operated on in the more recent years.[4]

Despite the reduction of mortality, the complication rate after extensive resection remains formidable. Pancreatic fistula, delayed gastric emptying, intraabdominal abscess are the most frequently reportedproblems. Extensive studies have been conducted to evaluate different measures to lower the development of pancreatic leak. The use of somatostatin or its analog,[5]application of fibrin glue around the pancreaticoenteric anastomosis,[6]supplemented by splinting the anastomosis with either an external[7,8]or internal stent,[9,10]putting the pancreatic remnant into the gastric lumen,[11,12]or alternatively, joining it to the jejunal loop by different methods of "dunking"[13-15]or placement of sutures,[16]including the use of operating microscope for accurate placement of fine sutures,[17]have all been reported with some success to reduce pancreatic fistula. As for the prevention of delayed gastric emptying, the use of intravenous erythromycin,[18]preservation of the left gastric vein,[19]construction of antecolic gastrojejunostomy together with a retro-gastric omental patch,[20]and performing an antrectomy instead of preserving the pylorus have all been explored.[21]One of the areas, which is less frequently addressed in the literature, is the potential contribution from vascular resection and reconstruction, either arterial or venous in origin, to the development of complications or even fatalities. The present review focuses primarily on the perioperative evaluation, and management of the associated vascular problems frequently encountered during the course of PD.

The celiac axis and superior mesenteric artery

In normal individuals, arterial stenosis from atherosclerosis or congenital anomalies is not infrequent. Although the blood supply from the celiac axis (CA), superior mesenteric artery (SMA), and branches of the hepatic artery might be compromised, there is no clinical importance under normal circumstances as there are rich interconnections between different arterial territories. According to the report by Song et al on 81 patients with CA stenosis and no aberrant hepatic artery, blood supply to the territories responsible by the CA derived from the SMA via either the pancreaticoduodenal arcades (95%) or the dorsal pancreatic artery (75%). When the aberrant right hepatic artery arises from the SMA, however, arterial communication from the intrahepatic interlobar collateral vessels, right gastric to left gastric arterial anastomosis, left hepatic to left gastric arterial anastomosis, peribiliary arterial plexuses are found instead.[22]As the gastroduodenal artery (GDA) and pancreatico-duodenal arcade are indubitary divided during the course of PD, the alternate pathway for blood supply to the organs covered by the CA is thus interrupted. While the liver and stomach are theoretically spared from an ischemic injury as they have additional route for blood supply, hepatico-jejunal and pancreatico-jejunal anastomoses are at risk of ischemia. Similar threats to the liver and the bilio-enteric anastomosis could exist if the replaced right hepatic artery has to be sacrificed as it courses through the back of the pancreatic head. A thorough understanding of the arterial anatomy of each patient, and careful planning of the operation, especially possible arterial reconstruction, could help to make PD a safer undertaking.

Incidence

The exact incidence of CA occlusion is uncertain from the literature. Even among normal individuals, CA stenosis was detected in 29 (7.3%) of the 400 asymptomatic Koreans, 16 (55%) of whom had compression from the medial arcuate ligament.[23]Thompson et al[24]reported that varying degree of CA occlusion could be found in 21 (10.5%) of 200 consecutive patients with no symptoms of visceral ischemia scheduled for PD. Patients with severe stenosis causing hemodynamic disturbance were reported. Trede[25]reported 6 patients (3%) of CA occlusion among 199 patients undergoing PDs. Kurosaki et al[26]found 5 (4%) of their 126 patients undergoing PD had clinically significant CA occlusion. In a recent French study, Gaujoux et al[27]detected 62 (11%) of their 545 patients who were prepared for PD had demonstrable visceral arterial occlusion; of whom 55 had their CA narrowed by the median arcuate ligament. Among these patients, 27% or 5% of the entire cohort reported had significant visceral arterial narrowing requiring therapeutic intervention.[27]The incidence of SMA occlusion, on the other hand, is seldom addressed, and was encountered in 3 (0.5 %) patients in the same French study.

Evaluation

Before the advent of computed tomography, angiography is often conducted as part of preoperative evaluation for selected patients before PD, especially in the 1980s. Visceral arteriography would help to certain extent, differentiate an inflammation from a neoplastic mass, an exocrine lesion from an islet cell tumor, determine the respectability of the tumor, and delineate the vascular anatomy.[24]Nowadays, such relatively invasive investigation has been largely replaced by multi-detector computed tomography, with 3-dimensional reconstruction of the vascular anatomy. In some centers, CT-reconstruction especially focusing on its lateral projection to examine the origins of the CA and SMA is an integral part of their preoperative evaluation process.[27]According to the causes of arterial obstruction, subsequent investigations differs. Thediagnosis of median arcuate ligament compression of the CA could be readily made on CT angiography as it has a characteristic "hook pattern" with a smooth cephalic indentation of the arterial lumen slightly distal to its junction with the aorta.[28]If atherosclerotic stenosis of one or both CA or SMA is suspected from CT, selective arteriography should be performed for detailed hemodynamic evaluation. Hemodynamically significant CA stenosis is considered if its branches are filled when contrast is selectively injected at the SMA via the gastroduodenal artery, pancreatico-duodenal arcades and hepatic artery. Similarly, significant SMA stenosis is established when it is visualized at selective injection of the CA and retrograde opacification through the same arterial arcade.

Treatment

In the presence of a significant occlusion to the CA and SMA, timing and the choice of treatment depend on the causes of arterial obstruction. Although preoperative work-up may be useful in revealing an ischemic threat, definitive assessment or recognition of the risk can only be made intra-operatively with a trial clamping of the GDA before its division.[29]Adequacy of hepatic inflow could be established by the strength of its pulsation, or better still, with the aid of intraoperative Doppler ultrasound for more objective evaluation to avoid a delayed recognition of an ischemic problem after pancreatectomy.[30]In the presence of a radiologically evident extrinsic compression of the celiac trunk, Berney et al[31]advocated a routine release of the arcuate ligament, even if GDA clamping does not alter blood flow, because it is easy to perform and involves no vascular anastomosis. On the other hand, when GDA clamping test is positive, the median arcuate ligament must be divided at beginning of the procedure, before GDA or pancreatic division. According to Gaujoux and associates, the "lateral" approach was their preference and was used in 19 of their 23 patients.[27]After a complete Kockerization of the duodenum, the right diaphragmatic crus was progressively divided on the right side of the abdominal aorta at the level of SMA origin. Thereafter, the right side and the upper edge of the CA were progressively freed of all dense fibrous tissue. An additional GDA clamping test and Doppler assessment were then repeated to confirm a complete release of the CA compression.

For patients with severe atherosclerotic stenosis of the CA and/or SMA, perioperative revascularization procedure must be performed if the pulse is stopped or decreased on trial clamping of the GDA, as otherwise fatal ischemic complication could occurred.[27]In patients with chronic pancreatitis, preservation of the GDA without compromising the oncological disease clearance could be considered as an option though technically dissection is difficult. In the presence of atheromatous occlusion with serious ischemic threat, a number of complex revascularization procedures are available, including bypass between the aorta and the CA,[32]between skeletonized splenic artery freed at the course of a total pancreatectomy and the SMA,[24]or between the aorta and the common hepatic artery.[33]Some preferred to put the vascular anastomosis away from pancreatico-jejunostomy by using vein graft to bridge the iliac artery and the splenic artery, so as to reduce the risk of exposure to pancreatic juice when a pancreatic fistula had developed.[34]The exact choice of arterial reconstruction used must be selected individually according to the intraoperative findings. Endovascular stenting is an attractive therapeutic option as stent placement could be performed at the same setting when patients with atherosclerotic occlusion of their CA or SMA are subjected to angiography for further hemodynamic evaluation. With preoperative endovascular stenting, all potential hazards associated with more complicated vascular bypass could be avoided.[35]Stent-assisted angioplasty is well accepted as a definitive treatment for stenotic major visceral arteries. In a series of 25 patients reported by Sharafuddin et al,[36]there were a success rate and a complication rate of 96% and 12%, respectively after the procedure. The 6-month patency rate in these patients was 92% and all patients were put on double anti-platelet agents. Nitnol stent is preferred for the CA as it could withstand the dynamic compression from the median arcuate ligament associated with respiration, and thus, runs a lesser risk of stent fracture.[36]Hasegawa et al[35]reported one patient with pancreatic cancer who had his stenotic CA stented at the time of selective angiography. After anticoagulant therapy for one week, the patient underwent an uneventful surgery 19 days after the endovascular procedure. Nonetheless, neither a median arcuate ligament division nor an extended nodal dissection around the CA was performed for the fear of dislodging the internal stent. Follow-up Doppler ultrasound performed three years after surgery showed sufficient arterial flow to the liver and CA.[35]Similar experiences were reported from France. Among the 7 patients with CT-evident atherosclerotic stenosis reported by Gaujoux et al,[27]3 of 4 patients (SMA 2 patients, CA 1) had endovascular stenting and the remaining patient with CA stenosis had intraoperative bypass. PD was performed 3 weeks to 4 months after the endovascular procedure according to suspected diagnosis.[27]

Replaced or acessory right hepatic artery

The classic report which described in details on the anatomical variations of the hepatic artery was published by Michels in 1966.[37]In 200 autopsy cases, anatomical variation was found in 45%. The right hepatic artery (type 3) and the common hepatic artery (type 9) were found to arise from the SMA in 18% and 2.5% of the cases, respectively. In another study on 1000 cases collected during the course of liver transplantation at UCLA, USA, Hiatt et al[38]reported that the corresponding incidence for the accessory or replaced right hepatic artery and the common hepatic artery stemming from the SMA was 10.6% and 1.5%, respectively. In a recent study by Turrini et al,[39]the accessory or replaced right hepatic artery was found in 10% of their 471 patients undergoing PD, of whom nearly half did not have a correct preoperative diagnosis as they were probably too small to detect even with thin-cut CT. Hence, acute awareness of the possibility and careful verification of all the arterial branches along the hepatoduodenal ligament during PD are essential to avoid unnecessary arterial injury. As the replaced or accessory right hepatic artery is behind the head of pancreas, their preservation could be difficult as invasion by tumor is not infrequent. On the other hand, their integrity is of paramount importance to ensure a safe bilio-enteric anastomosis.[40,41]Under such circumstances, preparation for arterial reconstruction, which often necessitates the use of micro-vascular techniques, should be made available.

Different options are available for the reconstruction of the replaced right hepatic artery after its resection. Selection of appropriate techniques depends on the specific situation developed after the resection. Although prosthetic and vein grafts could be chosen, autogenous grafts are always preferred as it involves no second incision and no risk of placing a foreign material in a potentially contaminated field. One useful technique is to preserve a sufficient length of GDA stump, and transposed to anastomose with the distal cut end of the replaced right hepatic artery along the hepatoduodenal ligament.[42]In the event when interposition vein graft is required, the use of gonadal vein is advocated as it has the advantage of a perfect size match with the hepatic artery, and is readily accessible without another incision.[43]with advanced disease. While a survival benefit derived from a routineen blocconcomitant resection of the retro-pancreatic PV is dubious,[44]major venous resection and reconstruction are often required to allow a complete macroscopic resection of malignant pancreatic head lesions. In a recent collective review of 6333 patients from 52 selected manuscripts provided an excellent background on the current status of PV/SMV resection during PD including total pancreatectomy. A total of 1646 (26%) patients, ranging from 2% to 77% according to individual studies, had concomitant PV/SMV resection and the median postoperative complication rate and mortality rate were 42% and 5.9%, respectively. Most (72.8%) of these patients had a circumferential PV/SMV resection and the median length resected was 3.9 cm (0.8-10).[45]

Portal venous system

As the portal vein (PV) courses behind the pancreatic neck, and the superior mesenteric vein (SMV) and its major tributaries lie over the anterior surface of the uncinate process, neoplastic involvement is common

Management of the portal vein

Depending on the site of tumor invasion of the PV, and thus, the extent of venous resection required, different technical options for resection and reconstruction are available. Tangential resection is possible when the lesion is adherent to a small part of the lateral or posterior wall of the PV and SMV. Repair is accomplished with vein patch harvested from the great saphenous vein or elsewhere if the venous lumen is significantly narrowed.[46]When a segment of the PV has to be sacrificed, primary end-to-end anastomosis should be made with preservation of all venous branches, including the splenic vein (SV), whenever feasible without using an interposition graft.

The resection and reconstruction of PV/SMV should be deferred until the rest of the operative specimen has been completely detached from all the surrounding structures including the SMA, and proximal and distal controls must be secured for the PV, SMV and splenic vein. The advantage is to avoid the need for venous anastomosis before the removal of the specimen, and minimize the time for venous occlusion. Concurrent inflow occlusion of the SMA at the same time as venous clamping is frequently employed to reduce small bowel edema, making subsequent biliary and pancreatic reconstruction less difficult. Systemic heparinization at the time of venous resection and reconstruction is not a routine practice, but when considered necessary, 2500 to 5000 U could be given at the time of clamping.[47]Anticoagulation after surgery is adequate with aspirin alone, but the patient should be put on heparin, followed by coumadin, if clot is found on postoperative imaging studies.[48]In some centers, anti-coagulation therapy, as guided by maintaining a partial thromboplastin time of 40 to 50 seconds, was limited to the initial 5 dayspostoperatively.[49]

Under normal circumstances, a loss of 2 cm or under of the PV/SMV needs no additional maneuver before an end-to-end venous anastomosis could be done without tension.[50]After all the soft tissue attachments are separated, leaving only the PV and SMV attaching to the specimen, the base of the small bowel mesentery is pushed cephalically to examine whether an approximation of the two cut ends of the veins is possible. When the presumed cut ends are judged to be too short for a tension-free venous anastomosis, the first step to overcome the problem is a complete mobilization of the right colon together with its mesocolon, and the root of the mesentery. In addition, the root of the transverse mesocolon is also detached from the anterior surface of the duodenum and pancreas. During PD, the ligament of Trietz, the third and fourth parts of the duodenum, and proximal jejunum should have been mobilized from the posterior abdominal wall. With these extensive mobilizations which allow the maximal cephalad displacement of the distal SMV stump, plus the division of the SV to improve the longitudinal mobility of the PV/SMV,[46]vein graft can be successfully spared even for a segmental loss of 7 to 10 cm.[51]If necessary, the falciform ligament, right coronary ligament and the right triangular ligament could be divided to allow placement of surgical packs to displace the proximal PV stump caudally to further ease the tension at the proposed venous anastomosis during construction.[52]On the other hand, some authors advocated the use of an interposition graft to bridge the ends of the PV/SMV instead of a routine sacrifice of the SV.[50]

Management of the splenic vein

The SV is divided when tumor invasion at its junction with the PV is evident, extra mobility for a direct end-toend anastomosis between the PV and SMV is necessary, and rarely nowadays, to provide better exposure for a thorough nodal clearance and soft tissue dissection at the proximal 3 to 4 cm of the SMA.[46,53]After division of the SV, a mandatory reconstruction is not a universally accepted practice. Some claimed that most patients would have no problem as the venous flow from the spleen and stomach could return to the systemic vein or the SMV through the short gastric vein and the esophageal vein.[54]In fact, recent studies on the anatomical relationship between the inferior mesenteric vein (IMV), SV and SMV using helical CT venography demonstrated that 48.3% to 68.5% of a normal population have their IMV joined to the SV, thus sparing the IMV/SV confluence when the SV is ligated closely to the PV.[53,55,56]On the other hand, a significant proportion of patients have their IMV joining to either the SMV (18.5% to 31%) or the junction between the SMV and SV (7.6% to 13.8%), and would run a potential risk of segmental left-side venous hypertension with resulted splenomegaly, hypertensive gastropathy, esophageal varices and hemorrhage with SV ligation. Alternative postulation assumed that leftside venous hypertension would develop only when the SV is diffusely thrombosed.[57]Using CT to evaluate the remnant SV patency after division close to its junction with the PV during PD 6 to 8 months after surgery, Strasberg et al[57]found that all 5 patients had a length of the SV remains patent, ranging 4.5 to 11 cm from the splenic hilum. Venous drainage from the spleen was shown to be channeled via multiple collaterals and none of them developed splenomegaly though the anatomical relationship of the SV and IMV was not described. Since the number of patients studied was small, they concluded that left-side venous hypertension is not an inevitable event after SV ligation without reconstruction.

Many had adopted a selective approach for SV reconstruction.[53,58,59]In the absence of an intact natural confluence between the SV and IMV, the two veins are anastomosed together with 8-0 non-absorbable sutures so that the venous drainage of the spleen and gastric remnant is preserved without making the more difficult anastomosis between the shortened SV and PV. Using surrogate markers including the changes of platelet count and spleen volume before and after surgery, Ferreira et al[58]found that such selective approach is feasible in obviating venous congestion during a shortterm follow-up. Re-implantation of the SV cut end to the side of an inter-position autologous vein graft has also been reported.[48]

Management of the superior mesenteric vein and its first order branches

The anatomy of the SMV varies in normal individuals. There are two anterior branches arising from the SMV, namely the middle colic vein and the gastro-colic trunk of Henle at the level of the root of the transverse mesocolon. The gastro-colic trunk drains several smaller tributaries including the right gastroepiploic vein, the right superior colic vein, and the antero-superior pancreatico-duodenal veins. Distal to the pancreas, there is a main trunk of the SMV, which is formed by the union of the ileal and jejunal branches in over 90% of normal individuals. The jejunal branch usually assumes a more horizontal path to go behind the SMA, but in an estimated 20% of the normal population their jejunal branches take a course anterior to the SMA instead. When the ileal and jejunal branches merge at the level of the SV without forming a mainSMV trunk, the jejunal branch would travels anterior to the SMA, and drains the IMV.[56,60]The relative position of the jejunal branch to the SMA bears relevance to the mobilization of the uncinate process from the SMV at PD. When the jejunal branch lies posterior to the SMA, it receives the small venous branches from the uncinate process. When the jejunal branch runs anterior to the SMA, the uncinate process has its venous drainage directed to the ileal branch, making it much easier to separate the uncinate process from the SMV.[61]

When the main trunk of the SMV and the junction of the ileal and jejunal branches are invaded by the tumor, segmental resection follow by reconstruction is necessary. As the small bowel would have adequate venous return if only one of the two major branches remains intact, the jejunal branch is usually sacrificed as it is usually posteriorly located with a thin wall, and difficult to access for the anastomosis to the SMV trunk.[46]While the ileal branch is always the preferred choice for reconstruction after segmental excision, additional consideration must also be given to its caliber. Based on the experiences gathered at the MD Anderson Cancer Center, an ileal branch of adequate caliber should have a diameter of 1.5 times larger than that of the SMA as seen on CT scan.[61]

Selection of vein substitution

Following segmental resection of the PV/SMV truck, interposition graft with different materials is occasionally required. The collective review on the 1198 patients with details on their venous reconstruction reflected best the general sentiments among surgeons. Primary anastomosis (88.6%) was used most frequently, followed by autologous vein graft (9.7%) and synthetic vein graft (1.7%).[45]The use of autologous graft is largely preferred over synthetic graft as the operative field after PD is potentially contaminated, especially if anastomotic leakage does occur.

While the preferred autologous vein varies according to individual centers, the accessibility of the vein graft, ease of procurement and the absence of long-term sequelae after its harvesting are the prime considerations for selection. The use of internal jugular vein,[46]superficial femoral vein[62]and IMV has been reported but gained limited enthusiasm. At the Mayo Clinic in Rochester, the left renal vein (LRV) is used to re-establish the venous continuity when a primary end-to-end anastomosis fails.[47]The LRV is harvested when the posterior retroperitoneal dissection has been completed leaving the specimen attaching only to the PV/SMV trunk. The LRV is generously exposed by further extending the Kockerization towards the left until its lateral limit, as defined by the insertion of the left gonadal vein and left adrenal vein to the LRV. The LRV is then divided with a mechanical linear stapler, taking care to protect both the left gonadal and left adrenal veins. The procedure is repeated on the right side flush with the inferior vena cava. The usual length obtained is about 3 to 4 cm but a length of 6 cm has been reported.[63]The vein segment is placed in heparinized saline until the time it is utilized. After harvesting, there is an initial transient rise in the serum creatinine level which returns to normal by the time of discharge.[48]

Reports on the outcome of artificial graft are sparse as most surgeons would use it as the last option. Theoretically, the use of polytetrafluoroethylene (PTFE) has the advantage that the external reinforcement ring could help to maintain a better patency when used in the high-flow, low-pressure and high-volume portal system. An anecdotal report made by Stauffer et al[64]showed encouraging results. Postoperatively estimated cumulative graft patency at one month for their 9 patients was 100%, as compared with 86%, and 60% after autologous vein and primary anastomotic repairs, respectively.[64]Retrospective experiences reported elsewhere suggested differently. In a series of 18 patients with PTFE reconstruction, the actual rate of thrombosis was 33%, as compared to a combined 12% in 13 primary end-to-end and 29 lateral venorrhaphy repairs. There was no statistically significant difference. None of the 4 patients in the same series who had used LRV for reconstruction developed graft thrombosis.[47]

The largest experience on PTFE graft reported today is a multi-center study on 33 patients from 3 United States institutions. Ring graft (73%) with a median diameter of 12 mm was used most often. Evaluation of patency commenced on the first postoperative day, and repeated on discharge. Aspirin or low-dose warfarin would be given for 3 to 6 months after surgery. Early graft thrombosis developed in 3 patients (14.2%). Two of them presented with gross ascites, and one died. Late thrombosis developed in another 5 patients, of whom 4 had recurrent disease. In contrast to early thrombosis, graft occlusion beyond 30 days was not associated with adverse clinical events. The actual rate of graft thrombosis in a mean follow-up of 14 months was 24.2%. Pancreatic fistula was found in 9 patients and no patient had graft infection. Estimated mean graft patency rate was 25.8 months. Ring reinforcement and large graft (over 12 mm) had no apparent benefit on the graft patency rate. In view of the comparable graft patency, the authors used PTFE more liberally in their practice especially for patients with good performance status and no gross abdominal contamination at operation.[65]

Conclusions

During PD, vascular resection and reconstruction, either arterial or venous, post additional technical problems to the already complex operative procedure. Adequate preoperative assessment of any patient planning for PD must include a CT which had been scrutinized by the surgical team involved. Examinations should focus not only the relationship between the lesions and the neighboring major vessels, but also a thorough elucidation of the vascular anatomy, such as the orifices of the celiac axis and SMA, the presence of replaced right hepatic artery and the relationship between the SMV and its major tributaries. Even if major arterial stenosis is not detected on preoperative CT, a trial clamping of the GDA, and confirmation of a good hepatic arterial flow before committing to proceed with a PD should be taken as a routine. Furthermore, careful palpation along the right distal border of the hepato-duodenal ligament, supplemented by intraoperative Doppler ultrasound, would be advisable to rule out the presence of a replaced or accessary right hepatic artery missed by CT.

If a segmental PV or SMV resection of 2 cm or more is required, a primary end-to-end venous anastomosis with preservation of the SV is the primary goal. In order to compensate for the loss of length and to allow a tension-free anastomosis, the first step should be a complete mobilization of the right mesocolon and small bowel mesentery. Division of the SV is the usual next move taken to gain more length. However, the necessity of a routine reconstruction after SV ligation has not been widely practiced irrespective of whether a natural confluence between the SV and IMV is present or not. If conservation of the spleno-portal venous junction is desired, inter-position graft to bridge the gap is the only choice. While autologous vein graft is often elected over synthetic graft, the use of left renal vein or internal jugular vein as the natural venous substitute had gained the most supports. Artificial vascular graft, especially ring-enforced PTFE graft, has received more attention in recent years. As the data on the infection rate, shortterm and long-term patency rate associated with these synthetic grafts had been scanty, their application should remain cautious at present.

Contributors: LECS proposed the study, performed research and wrote the first draft. LECS is the guarantor.

Funding: None.

Ethical approval: Not needed.

Competing interest: No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

1 Whipple AO, Parsons WB, Mullins CR. Treatment of carcinoma of the ampulla of Vater. Ann Surg 1935;102:763-779.

2 Trede M, Schwall G, Saeger HD. Survival after pancreatoduodenectomy. 118 consecutive resections without an operative mortality. Ann Surg 1990;211:447-458.

3 Winter JM, Cameron JL, Campbell KA, Arnold MA, Chang DC, Coleman J, et al. 1423 pancreaticoduodenectomies for pancreatic cancer: A single-institution experience. J Gastrointest Surg 2006;10:1199-1211.

4 Mayo SC, Gilson MM, Herman JM, Cameron JL, Nathan H, Edil BH, et al. Management of patients with pancreatic adenocarcinoma: national trends in patient selection, operative management, and use of adjuvant therapy. J Am Coll Surg 2012;214:33-45.

5 Connor S, Alexakis N, Garden OJ, Leandros E, Bramis J, Wigmore SJ. Meta-analysis of the value of somatostatin and its analogues in reducing complications associated with pancreatic surgery. Br J Surg 2005;92:1059-1067.

6 Lillemoe KD, Cameron JL, Kim MP, Campbell KA, Sauter PK, Coleman JA, et al. Does fibrin glue sealant decrease the rate of pancreatic fistula after pancreaticoduodenectomy? Results of a prospective randomized trial. J Gastrointest Surg 2004;8:766-774.

7 Poon RT, Fan ST, Lo CM, Ng KK, Yuen WK, Yeung C, et al. External drainage of pancreatic duct with a stent to reduce leakage rate of pancreaticojejunostomy after pancreaticoduodenectomy: a prospective randomized trial. Ann Surg 2007;246:425-435.

8 Pessaux P, Sauvanet A, Mariette C, Paye F, Muscari F, Cunha AS, et al. External pancreatic duct stent decreases pancreatic fistula rate after pancreaticoduodenectomy: prospective multicenter randomized trial. Ann Surg 2011;253:879-885.

9 Winter JM, Cameron JL, Campbell KA, Chang DC, Riall TS, Schulick RD, et al. Does pancreatic duct stenting decrease the rate of pancreatic fistula following pancreaticoduodenectomy? Results of a prospective randomized trial. J Gastrointest Surg 2006;10:1280-1290.

10 Smyrniotis V, Arkadopoulos N, Kyriazi MA, Derpapas M, Theodosopoulos T, Gennatas C, et al. Does internal stenting of the pancreaticojejunostomy improve outcomes after pancreatoduodenectomy? A prospective study. Langenbecks Arch Surg 2010;395:195-200.

11 McKay A, Mackenzie S, Sutherland FR, Bathe OF, Doig C, Dort J, et al. Meta-analysis of pancreaticojejunostomy versus pancreaticogastrostomy reconstruction after pancreaticoduodenectomy. Br J Surg 2006;93:929-936.

12 Wente MN, Shrikhande SV, Müller MW, Diener MK, Seiler CM, Friess H, et al. Pancreaticojejunostomy versus pancreaticogastrostomy: systematic review and meta-analysis. Am J Surg 2007;193:171-183.

13 Berger AC, Howard TJ, Kennedy EP, Sauter PK, Bower-Cherry M, Dutkevitch S, et al. Does type of pancreaticojejunostomy after pancreaticoduodenectomy decrease rate of pancreatic fistula? A randomized, prospective, dual-institution trial. J Am Coll Surg 2009;208:738-749.

14 Peng SY, Wang JW, Lau WY, Cai XJ, Mou YP, Liu YB, et al. Conventional versus binding pancreaticojejunostomy after pancreaticoduodenectomy: a prospective randomized trial. Ann Surg 2007;245:692-698.

15 Buc E, Flamein R, Golffier C, Dubois A, Nagarajan G,Futier E, et al. Peng's binding pancreaticojejunostomy after pancreaticoduodenectomy: a French prospective study. J Gastrointest Surg 2010;14:705-710.

16 Grobmyer SR, Kooby D, Blumgart LH, Hochwald SN. Novel pancreaticojejunostomy with a low rate of anastomotic failure-related complications. J Am Coll Surg 2010;210:54-59.

17 Hashimoto Y, Traverso LW. Pancreatic anastomotic failure rate after pancreaticoduodenectomy decreases with microsurgery. J Am Coll Surg 2010;211:510-521.

18 Yeo CJ, Barry MK, Sauter PK, Sostre S, Lillemoe KD, Pitt HA, et al. Erythromycin accelerates gastric emptying after pancreaticoduodenectomy. A prospective, randomized, placebocontrolled trial. Ann Surg 1993;218:229-238.

19 Kurosaki I, Hatakeyama K. Preservation of the left gastric vein in delayed gastric emptying after pylorus-preserving pancreaticoduodenectomy. J Gastrointest Surg 2005;9:846-852.

20 Nikfarjam M, Kimchi ET, Gusani NJ, Shah SM, Sehmbey M, Shereef S, et al. A reduction in delayed gastric emptying by classic pancreaticoduodenectomy with an antecolic gastrojejunal anastomosis and a retrogastric omental patch. J Gastrointest Surg 2009;13:1674-1682.

21 Diener MK, Knaebel HP, Heukaufer C, Antes G, Büchler MW, Seiler CM. A systematic review and meta-analysis of pyloruspreserving versus classical pancreaticoduodenectomy for surgical treatment of periampullary and pancreatic carcinoma. Ann Surg 2007;245:187-200.

22 Song SY, Chung JW, Kwon JW, Joh JH, Shin SJ, Kim HB, et al. Collateral pathways in patients with celiac axis stenosis: angiographic-spiral CT correlation. Radiographics 2002;22: 881-893.

23 Park CM, Chung JW, Kim HB, Shin SJ, Park JH. Celiac axis stenosis: incidence and etiologies in asymptomatic individuals. Korean J Radiol 2001;2:8-13.

24 Thompson NW, Eckhauser FE, Talpos G, Cho KJ. Pancreaticoduodenectomy and celiac occlusive disease. Ann Surg 1981;193:399-406.

25 Trede M. The surgical treatment of pancreatic carcinoma. Surgery 1985;97:28-35.

26 Kurosaki I, Hatakeyama K, Nihei KE, Oyamatsu M. Celiac axis stenosis in pancreaticoduodenectomy. J Hepatobiliary Pancreat Surg 2004;11:119-124.

27 Gaujoux S, Sauvanet A, Vullierme MP, Cortes A, Dokmak S, Sibert A, et al. Ischemic complications after pancreaticoduodenectomy: incidence, prevention, and management. Ann Surg 2009;249:111-117.

28 Horton KM, Talamini MA, Fishman EK. Median arcuate ligament syndrome: evaluation with CT angiography. Radiographics 2005;25:1177-1182.

29 Murakami Y, Uemura K, Yokoyama Y, Sasaki M, Morifuji M, Hayashidani Y, et al. Celiac axis occlusion with replaced common hepatic artery and pancreatoduodenectomy. J Gastrointest Surg 2004;8:520-522.

30 Nara S, Sakamoto Y, Shimada K, Sano T, Kosuge T, Takahashi Y, et al. Arterial reconstruction during pancreatoduodenectomy in patients with celiac axis stenosis--utility of Doppler ultrasonography. World J Surg 2005;29:885-889.

31 Berney T, Pretre R, Chassot G, Morel P. The role of revascularization in celiac occlusion and pancreatoduodenectomy. Am J Surg 1998;176:352-356.

32 Takach TJ, Livesay JJ, Reul GJ Jr, Cooley DA. Celiac compression syndrome: tailored therapy based on intraoperative findings. J Am Coll Surg 1996;183:606-610.

33 Miyata M, Takao T, Okuda A, Sasako Y, Sunada S. Pancreatoduodenectomy for periampullary cancer associated with celiac occlusion: a case report. Surgery 1988;103:261-263.

34 Okamoto H, Suminaga Y, Toyama N, Konishi F, Kawahito H. Autogenous vein graft from iliac artery to splenic artery for celiac occlusion in pancreaticoduodenectomy. J Hepatobiliary Pancreat Surg 2003;10:109-112.

35 Hasegawa K, Imamura H, Akahane M, Miura Y, Takayama T, Ohtomo K, et al. Endovascular stenting for celiac axis stenosis before pancreaticoduodenectomy. Surgery 2003;133:440-442.

36 Sharafuddin MJ, Olson CH, Sun S, Kresowik TF, Corson JD. Endovascular treatment of celiac and mesenteric arteries stenoses: applications and results. J Vasc Surg 2003;38:692-698.

37 Michels NA. Newer anatomy of the liver and its variant blood supply and collateral circulation. Am J Surg 1966;112:337-347.

38 Hiatt JR, Gabbay J, Busuttil RW. Surgical anatomy of the hepatic arteries in 1000 cases. Ann Surg 1994;220:50-52.

39 Turrini O, Wiebke EA, Delpero JR, Viret F, Lillemoe KD, Schmidt CM. Preservation of replaced or accessory right hepatic artery during pancreaticoduodenectomy for adenocarcinoma: impact on margin status and survival. J Gastrointest Surg 2010;14:1813-1819.

40 Traverso LW, Freeny PC. Pancreaticoduodenectomy. The importance of preserving hepatic blood flow to prevent biliary fistula. Am Surg 1989;55:421-426.

41 Hamazaki K, Mimura H, Kobayashi T, Kim H, Sakagami K, Orita K. Hepatic artery reconstruction after resection of the hepatoduodenal ligament. Br J Surg 1991;78:1366-1367.

42 Allendorf JD, Bellemare S. Reconstruction of the replaced right hepatic artery at the time of pancreaticoduodenectomy. J Gastrointest Surg 2009;13:555-557.

43 Sarmiento JM, Panneton JM, Nagorney DM. Reconstruction of the hepatic artery using the gastroduodenal artery. Am J Surg 2003;185:386-387.

44 Fortner JG. Regional resection of cancer of the pancreas: a new surgical approach. Surgery 1973;73:307-320.

45 Siriwardana HP, Siriwardena AK. Systematic review of outcome of synchronous portal-superior mesenteric vein resection during pancreatectomy for cancer. Br J Surg 2006;93:662-673.

46 Tseng JF, Raut CP, Lee JE, Pisters PW, Vauthey JN, Abdalla EK, et al. Pancreaticoduodenectomy with vascular resection: margin status and survival duration. J Gastrointest Surg 2004;8:935-950.

47 Smoot RL, Christein JD, Farnell MB. Durability of portal venous reconstruction following resection during pancreaticoduodenectomy. J Gastrointest Surg 2006;10:1371-1375.

48 Smoot RL, Christein JD, Farnell MB. An innovative option for venous reconstruction after pancreaticoduodenectomy: the left renal vein.J Gastrointest Surg 2007;11:425-431.

49 Riediger H, Makowiec F, Fischer E, Adam U, Hopt UT. Postoperative morbidity and long-term survival after pancreaticoduodenectomy with superior mesenterico-portal vein resection.J Gastrointest Surg 2006;10:1106-1115.

50 Cusack JC Jr, Fuhrman GM, Lee JE, Evans DB. Managing unsuspected tumor invasion of the superior mesentericportal venous confluence during pancreaticoduodenectomy. Am J Surg 1994;168:352-354.

51 Fujisaki S, Tomita R, Fukuzawa M. Utility of mobilizationof the right colon and the root of the mesentery for avoiding vein grafting during reconstruction of the portal vein. J Am Coll Surg 2001;193:576-578.

52 Zhang J, Qian HG, Leng JH, Cui M, Qiu H, Zhou GQ, et al. Long mesentericoportal vein resection and end-to-end anastomosis without graft in pancreaticoduodenectomy. J Gastrointest Surg 2009;13:1524-1528.

53 Misuta K, Shimada H, Miura Y, Kunihiro O, Kubota T, Endo I, et al. The role of splenomesenteric vein anastomosis after division of the splenic vein in pancreatoduodenectomy. J Gastrointest Surg 2005;9:245-253.

54 Fortner JG. Technique of regional subtotal and total pancreatectomy. Am J Surg 1985;150:593-600.

55 Sakaguchi T, Suzuki S, Morita Y, Oishi K, Suzuki A, Fukumoto K, et al. Analysis of anatomic variants of mesenteric veins by 3-dimensional portography using multidetector-row computed tomography. Am J Surg 2010;200:15-22.

56 Graf O, Boland GW, Kaufman JA, Warshaw AL, Fernandez del Castillo C, Mueller PR. Anatomic variants of mesenteric veins: depiction with helical CT venography. AJR Am J Roentgenol 1997;168:1209-1213.

57 Strasberg SM, Bhalla S, Sanchez LA, Linehan DC. Pattern of venous collateral development after splenic vein occlusion in an extended Whipple procedure : comparison with collateral vein pattern in cases of sinistral portal hypertension. J Gastrointest Surg 2011;15:2070-2079.

58 Ferreira N, Oussoultzoglou E, Fuchshuber P, Ntourakis D, Narita M, Rather M, et al. Splenic vein-inferior mesenteric vein anastomosis to lessen left-sided portal hypertension after pancreaticoduodenectomy with concomitant vascular resection. Arch Surg 2011;146:1375-1381.

59 Tamura K, Sumi S, Koike M, Yano S, Nagami H, Nio Y. A splenic-inferior mesenteric venous anastomosis prevents gastric congestion following pylorus preserving pancreatoduodenectomy with extensive portal vein resection for cancer of the head of the pancreas. Int Surg 1997;82:155-159.

60 Kim HJ, Ko YT, Lim JW, Lee DH. Radiologic anatomy of the superior mesenteric vein and branching patterns of the first jejunal trunk: evaluation using multi-detector row CT venography. Surg Radiol Anat 2007;29:67-75.

61 Katz MH, Fleming JB, Pisters PW, Lee JE, Evans DB. Anatomy of the superior mesenteric vein with special reference to the surgical management of first-order branch involvement at pancreaticoduodenectomy. Ann Surg 2008;248:1098-1102.

62 Fleming JB, Barnett CC, Clagett GP. Superficial femoral vein as a conduit for portal vein reconstruction during pancreaticoduodenectomy. Arch Surg 2005;140:698-701.

63 Miyazaki M, Itoh H, Kaiho T, Ambiru S, Togawa A, Sasada K, et al. Portal vein reconstruction at the hepatic hilus using a left renal vein graft. J Am Coll Surg 1995;180:497-498.

64 Stauffer JA, Dougherty MK, Kim GP, Nguyen JH. Interposition graft with polytetrafluoroethylene for mesenteric and portal vein reconstruction after pancreaticoduodenectomy. Br J Surg 2009;96:247-252.

65 Chu CK, Farnell MB, Nguyen JH, Stauffer JA, Kooby DA, Sclabas GM, et al. Prosthetic graft reconstruction after portal vein resection in pancreaticoduodenectomy: a multicenter analysis. J Am Coll Surg 2010;211:316-324.

April 18, 2012

Accepted after revision May 16, 2012

Author Affiliations: Pedder Clinic, 4/F 9 Queen's Road Central, Hong Kong SAR, China (Lai ECS)

Edward CS Lai, MS, Pedder Clinic, 4/F 9 Queen's Road Central, Hong Kong SAR, China (Tel: 852-28689333; Fax: 852-28689223; Email: ecslai@hku.hk)

? 2012, Hepatobiliary Pancreat Dis Int. All rights reserved.

10.1016/S1499-3872(12)60154-4