Superior mesenteric artery first approach versus standard pancreaticoduodenectomy: a systematic review and meta-analysis

Ionut Negoi， Sorin Hostiuc， Alexandru Runcanu， Ruxandra Irina Negoi and Mircea Beuran

Bucharest， Romania

Superior mesenteric artery first approach versus standard pancreaticoduodenectomy: a systematic review and meta-analysis

Ionut Negoi， Sorin Hostiuc， Alexandru Runcanu， Ruxandra Irina Negoi and Mircea Beuran

Bucharest， Romania

BACKGROUND: The superior mesenteric artery (SMA) first approach was proposed recently as a new modification of the standard pancreaticoduodenectomy. Increasing evidence showed that a periadventiceal dissection of the SMA with early transection of the inflow during pancreaticoduodenectomy associates better early perioperative results， and setup the scene for long-term oncological benefits. The objectives of the current study are to compare the operative results and long-term oncological outcomes of SMA first approach pancreaticoduodenectomy (SMA-PD) with standard pancreaticoduodenectomy (S-PD).

DATA SOURCES: Electronic search of the PubMed/MEDLINE， EMBASE， Web of Science and Cochrane Library was performed until July 2015. We considered randomized controlled trials (RCTs) and non-randomized comparative studies (NRCSs) comparing SMA-PD with S-PD to be eligible if they included patients with periampullary cancers.

RESULTS: A total of one RCT and thirteen NRCSs met the inclusion criteria， involving 640 patients with SMA-PD and 514 patients with S-PD. The SMA-PD was associated with less intraoperative bleeding， less blood transfusions and higher rate of associated venous resections. The pancreatic fistula and delayed gastric emptying had a significantly lower rate in the SMA-PD group. There were no differences between the two approaches regarding overall complications， major complication rates and in-hospital mortality. There was no difference regarding R0 resection rate， and one-， two- or three-year overall survival. The SMA-PD was associated with a lower local， hepatic and extrahepatic metastatic rate.

CONCLUSIONS: The SMA-PD is associated with better perioperative outcomes， such as blood loss， transfusion requirements， pancreatic fistula， and delayed gastric emptying. Although the one-， two- or three-year overall survival rate is not superior， the SMA-PD has a lower local and metastatic recurrence rate.

(Hepatobiliary Pancreat Dis Int 2017;16:127-138)

pancreaticoduodenectomy;

superior mesenteric artery; artery first;

cancer

Introduction

Surgical resection represents the only potentially curative approach for patients with pancreatic cancer.[1]Due to the aggressive biology of this malignancy， less than 20% of patients can be resected， and even under these optimal clinical conditions the median survival ranges were from 20.1 to 23.6 months.[2]In an effort to improve the local control of the disease， different surgical modifications of the standard Whipple technique，[3]like extended lymphadenectomy[4]or regional pancreatectomy，[5]have been proposed. Unfortunately none of them was associated with improved long-term oncological benefits， while retaining a significant associated morbidity. During the recent years， a new modification of the standard pancreaticoduodenectomy was proposed， concept termed superior mesenteric artery (SMA) first approach.[6]In a standard pancreatoduodenectomy the pancreatic neck is transected， and the specimen is dissected from the superior mesenteric (SMV)-portal vein (PV) complex， ligating all the small veins draining the pancreatic head.[7]The final step involves separation of the pancreatic head from the SMA. In the SMA first approach pancreaticoduodenectomy the inferior pancre-aticoduodenal artery (IPDA) and the periadventiceal dissection of the artery is performed without dividing the pancreatic neck.[6，8]Thus the tumor infiltration of the arterial wall， an absolute contraindication for resection is early detected， before the “point of no return’’.

The multi-detector computed tomography， preferred for preoperative staging，[9]has an accuracy for resectability of 95%， but this significantly decreases after neoadjuvant radiotherapy to a positive predictive value for unresectability of 25%.[10，11]For these patients， with borderline resectable tumors undergoing neoadjuvant therapy， early dissection of the SMA to determine its tumor infiltration is of particular importance. Recent evidence coming from high volume centers proved that 61%-85% of pancreatic cancer resections are R1 resections.[12-14]According to the National Comprehensive Cancer Network guideline the survival benefits of a R1 resection may be comparable to definitive chemoradiation without surgery.[2]The most frequent site of R1 resection is at the level of the mesopancreas，[15，16]represented by the retropancreatic retroportal tissue， on the right side of the SMA and celiac artery， through which runs the pancreaticoduodenal arteries， veins， lymphatics and nerve plexus.[17]Thus， a surgical step focused on periadventitial dissection of the SMA， in a bloodless field， at the beginning of the surgery may increase the total mesopancreas excision rate， decreasing the microscopic involvement of the retroperitoneal margin.

Initial dissection of the SMA may be indicated also for patients with locally advanced tumors， in those 11%-65% of cases when resection of the PV， SMV and/or splenic vein is required.[18]In this setting， early retropancreatic dissection is associated with a reduced blood loss and shorter PV clamping times.[18]

The objective of this systematic review and metaanalysis is to summarize the current evidence regarding early dissection of the SMA during pancreaticoduodenectomy (SMA-PD) and to compare its operative results and long-term oncological outcomes with the standard approach pancreaticoduodenectomy (S-PD).

Methods

This systematic review and meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.[19]Electronic search， study selection， data extraction and quality assessment were performed independently by three reviewers (NI， HS and NRI).

Search strategy

We conducted electronic search to identify all published randomized controlled trials (RCTs) and nonrandomized comparative studies (NRCSs)， using the following databases: PubMed/MEDLINE， EMBASE， Web of Science， Cochrane Center Register of Controlled Trails (CENTRAL)， Cochrane Database of Systematic Reviews (CDSR)， and Database of Abstracts of Reviews of Effectiveness (DARE)， since their beginning until July 2015. We use English， Spanish and French as language restrictions. The most recent search in PubMed was performed in July 2015.

We constructed the search strategy by using various combinations of the terms related SMA-PD or S-PD for periampullary tumors. We have used in different combinations of the following key words: “pancreaticoduodenectomy”， “superior mesenteric artery”， “artery first”， “pancreas”， and “cancer”. These were identified as truncated words in the title， abstract or in medical subject headings (MeSH). Electronic and manual cross-referencing was further used to find other relevant sources. The search strategy used in PubMed/MEDLINE was: [pancreaticoduodenectomy (MeSH Terms)] OR [pancreatoduodenectomy (Title/Abstract)] OR [duodenopancreatectomy (Title/Abstract)] AND [superior mesenteric artery (Title/Abstract)] OR [inferior pancreaticoduodenal artery (Title/Abstract)] OR [inferior pancreatoduodenal artery (Title/Abstract)] OR [uncinate process (Title/Abstract)] OR [mesopancreas (Title/Abstract)] OR [no-touch (Title/Abstract)].

Trial selection

We considered RCTs and NRCSs comparing SMA-PD with S-PD to be eligible if they included patients with periampullary cancers. An SMA first approach was considered when the dissection of the SMA with transection of the IPDA was performed prior to the separation of the specimen from the SMV-PV confluent.

Outcome measures

Primary outcome: one-， two-， and three-year overall survival. Secondary outcomes: quality of the resected specimen (R0 resection rate， lymph nodes retrieved)， surgical data (operation time， blood loss， transfusion requirement， percentage of PV resection， percentage of T3/T4 tumors)， postoperative complications (overall complications， major complications， pancreatic fistula， diarrhea)， intraoperative complications， length of hospital stay， 30-day mortality， rate of adjuvant chemotherapy， and recurrence of the disease.

Data extraction

Three authors (NI， HS and NRI) of the present review assessed the methodological quality of eligible trials and independently extracted data from individual studies，using a data extraction form. We extracted the following data: first author， year of publication， title， source， contact address， criteria for patient inclusion and exclusion， sample size， baseline characteristics which included mean age， gender ratio， location of the tumor， pathology data， number of patients assigned to each treatment group， details of intervention regimens， and all the outcomes of interest.

Assessment of methodological quality

To assess the risk of bias in RCTs we used the Cochrane Collaboration tool， which grade the random sequence generation， allocation concealment， blinding of participant and personnel， blinding of outcome assessment， incomplete outcome data， selective reporting and other biases.[20]For evaluation of NRCSs， we used the Cochrane Risk of Bias Assessment Tool for Non-Randomized Studies of Interventions (ACROBAT-NRSI) which grade bias due to confounding， bias in selection of participants into the study， bias in measurement of interventions， bias due to departures from intended interventions， bias due to missing data， bias in measurement of outcomes， bias in selection of the reported result.[21]

Statistical analysis

We used Review Manager Software 5.3.5 (The Cochrane Collaboration) and Comprehensive Meta-Analysis (CMA) Software Version 2 to analyze the data. Mean difference (MD) was chosen as effect measure for continuous data and odds ratio (OR) for dichotomous data， reported along with the 95% confidence interval (CI). In case of continuous data presented as median and range， we estimated the mean and standard deviation according to the methods described by Hozo et al.[22]We performed a subgroup analysis for patients with pancreatic ductal adenocarcinoma. Chi-square and I2statistics were used for assessing studies’ heterogeneity. I2value ≤25% indicate a lower heterogeneity， I2value >25% but ≤75% indicate a moderate heterogeneity and I2value >75% indicate higher heterogeneity.[23]Reasons for statistical heterogeneity were explored using sensitivity analyses， through exclusion of specific studies. We used a fixed-effect model analysis for outcomes with low heterogeneity. If present clinical heterogeneity between included studies， due to differences with respect to eligibility criteria (study population)， type of surgical technique， lacking or differing definitions of outcomes， meta-analysis were performed applying a random-effect model (DerSimonian-Laird method).[24]Begg’s funnel plot and Egger’s test were used for assessing publication bias.[25]The statistical significance was defined as P<0.1 in Egger’s test and P<0.05 for the other statistical tests. To correct possible publication bias， we performed trim and fill analysis.

Results

Description of studies

Results of the search

The initial electronic and manual literature searches revealed 219 full text articles. A total of one RCT (from Europe)，[26]and thirteen NRCSs (seven from Japan， two from India， two from Europe， one from Australia and one from China)[27-39]met the inclusion criteria for the qualitative and quantitative (meta-analysis) synthesis， involving 640 patients in SMA-PD and 514 in S-PD. In the SMA-PD group， there were 360 patients with pancreatic cancer and 266 patients in S-PD. The results of thirteen studies were published in English and one in Spanish. The reasons for exclusion at each stage of the process are shown in Fig. 1.

Included studies

The characteristics of the included studies are summarized in Table 1. All studies were published between 2007-2015. The sample size ranged from 12 to 287 patients. Four studies included only patients with pancreatic cancer.[26，32，33，37]All studies had a similar exclusion criterion， which was stage IV disease. Neoadjuvant therapy was performed in none of the studies. The perioperative care was not described in most studies. Patient demographics and baseline clinical data were similar between the SMAPD and S-PD groups， with a mean age of 68.5 vs 68.9 years and 60.8% vs 57.6% male patients， respectively.

Risk of bias in included studies

The risk of bias in the RCT[26]according to the review authors’ judgment was low for allocation concealment， blinding participants and personnel， blinding of outcome assessment， incomplete outcome data， and selective reporting domains. The risk of bias was considered unclear in the other bias domain. The risk of bias in thenon-randomized studies， including one prospective，[27]four from a prospective database，[28-31]and three cohort studies[32-34]were judged by the review authors’ as moderate. The remaining five case-control studies[35-39]had the lowest methodological quality (serious risk of bias) (Table 2).

Fig. 1. Flow diagram of the systematic literature search and study selection according to the PRISMA statement.

Table 1. The characteristics of the included studies

Table 2. Quality assessment of the included non-randomized studies using Cochrane Collaboration tools

Techniques to approach the SMA during SMA-PD

The posterior SMA first approach， used in four of the included studies，[29，30，36，39]assumes a wide Kocher maneuver with SMA’s origin dissection just above the left renal vein. The right side of the SMA is further distally skeletonized， with step-by-step transection of the retroportal lamina.[6，8]To overcome its technique challenges， especially in obese patients， the group from Strasbourg later introduced a combined approach， posterior and anterior with hanging maneuver.[40，41]Although an SMAPD requires refined surgical skills， the current metaanalysis showed that， in experienced hands， there are no differences for duration of surgery comparing with standard technique.

Anterior approach， used in four of the included studies，[26，28，31，35]is proposed as a no-touch isolationtechnique.[42]After isolation of the pancreatic neck， the posterolateral aspect of the SMV-PV is dissected. Then a hanging maneuver is practical， passing bluntly a tape on the anterior surface of the aorta， on the right side of the celiac and SMAs. Lifting this tape may allow dissection of the retroperitoneal margin of the specimen. Reversed Kocherisation is the latest step of the resection.[42]

During mesenteric approach， used in two of the included studies，[32，38]the SMA is approached from the left side of the duodenojejunal flexure. Starting from its origin， the posterior and right aspect of the SMA will be dissected over a few centimeters. The next step is identification of the SMA below the tumor， below the transverse mesocolon. Then， if possible from the inframesocolic position， the SMA is followed all the way downward toward its origin at the aorta.[43]

Left posterior approach， used in two of the included studies，[33，37]the superior mesenteric pedicle is dissected from the left side of the SMA， opening the retroperitoneum at the level of left duodenomesocolic fold.[37]The origin of the SMA is identified as the endpoint of dissection. SMA is isolated above the third segment of the duodenum， and skeletonized in a longitudinal direction from the origin of the middle colic artery up to the confluence with the aorta.[33]Further traction the jejunum to left rotates the SMA in a counterclockwise direction， allowing the division of the IPDA， and then the dissection of the SMV， located inferiorly to the SMA.[37]

For an uncinate process first approach， used in one of our studies，[27]the pancreatic head is retrograde dissected， from caudal to cranial.[44]After transection and mobilization， the first jejunal loop is transposed towards the right aspect of the upper abdomen. The specimen， together with the uncinate process and the pancreatic head is mobilized from retroperitoneal soft tissue， under clear visualization of the SMV and SMA. Transection ofthe pancreas represents the last step of the resection.[44]

Table 3. Results of a meta-analysis comparing superior mesenteric artery first approach pancreaticoduodenectomy with standard pancreaticoduodenectomy in periampullary tumors

Fig. 2. Meta-analysis of studies on transfusion requirements of patients undergoing superior mesenteric artery first approach pancreaticoduodenectomy (SMA-PD) versus standard pancreaticoduodenectomy (S-PD).

Effects of intervention

Intraoperative outcomes

Duration of surgery was reported in thirteen studies，[26-32，34-39]with 621 patients in the SMA-PD group and 495 patients in the S-PD group. The observed heterogeneity between studies was high. There was no statistical significant MD between SMA-PD and S-PD (MD=-9.94; 95% CI: -57.44 to 37.56 minutes; P=0.68; Table 3).

The blood loss was reported by twelve studies，[27-32，34-39]with 615 patients in the SMA-PD group and 489 patients in the S-PD group. There was a high heterogeneity between studies. The SMA-PD was associated with an MD of 345.34 mL less intraoperative bleeding (MD=-345.34; 95% CI: -477.09 to -213.59 mL; P<0.01).

Transfusion requirements were reported by eight studies，[26，27，29，31，32，34-36]with 430 patients in the SMA-PD group and 324 patients in the S-PD group. The heterogeneity between studies was high. Patients from SMA-PD required significantly less blood transfusions (OR=0.15; 95% CI: 0.06 to 0.37; P<0.01; Fig. 2).

The rate of SMV-PV resections was reported by ten studies，[28-33，35-37，39]with low heterogeneity between them. Patients from the SMA-PD group had a significantly higher rate of associated venous resections (OR=1.62; 95% CI: 1.09 to 2.39; P=0.02).

The rate of T3 and T4 tumors was reported by six studies，[26，30，34-37]with low heterogeneity between them. There was no statistical significant difference regarding the rate of locally advanced tumors between SMA-PD and S-PD (OR=1.13; 95% CI: 0.78 to 1.63; P=0.53).

Postoperative morbidity and mortality

Eleven studies[27，29-37，39]presented the overall morbidity， including 534 patients in the SMA-PD group and 407 in the S-PD group. There was a high statistical heterogeneity between the studies. There was no difference regarding overall complication rate between SMA-PD and S-PD (OR=0.81; 95% CI: 0.50 to 1.31; P>0.05).

Five studies[26，28-30，35]reported the rate of major complications， grade III-IV according to Clavien-Dindo scale， with a low statistical heterogeneity between studies. There was no difference between SMA-PD and S-PD (OR=1.09; 95% CI: 0.65 to 1.82; P>0.05).

The rate of pancreatic fistula was lower after SMA-PD (OR=0.54; 95% CI: 0.37 to 0.78; P=0.0009). This outcome was reported by twelve studies，[26-28，30-37，39]with a low heterogeneity between them. There were 560 patients in the SMA-PD and 431 patients in the S-PD group.

The rate of postoperative diarrhea requiring medication was significantly higher in the SMA-PD group (OR=1.91; 95% CI: 1.08 to 3.40; P=0.03). Four studies[28，32，33，37]reported this outcome， with a low heterogeneity between them.

The delayed gastric emptying had a significantly lower rate in the SMA-PD group (OR=0.42; 95% CI: 0.24 to 0.75; P=0.003). There was a moderate heterogeneity between the seven studies[28，31-34，36，37]with 313 patients in the SMA-PD group and 263 in the S-PD group.

The intraabdominal hemorrhage was reported by four studies，[31，36，37，39]with a low heterogeneity between studies. There was no difference between the two approaches (OR=1.16; 95% CI: 0.31 to 4.29; P>0.05).

Five studies[33，35-37，39]reported the surgical reintervention rate， with 293 patients in the SMA-PD group and 205 in the S-PD group. The statistical heterogeneity was low. There was no difference between the two approaches (OR=0.47; 95% CI: 0.21 to 1.04; P>0.05).

In-hospital mortality was reported by thirteen studies，[26-37，39]with 622 patients in the SMA-PD group and 496 patients in the S-PD group. There was a low heterogeneity between the studies and no statistical significant difference between the two surgical approaches (OR=0.47; 95% CI: 0.18 to 1.23; P>0.05).

The hospital stay was shorter for SMA-PD， with anMD of 2.52 days (MD=-2.52; 95% CI: -5.00 to -0.05; P=0.05)[.26-T30h， 3e2r

， 3e4， 3w6， 3a7s，39a]high heterogeneity between the studies.

Quality of oncological resection

Nine studies[26，27，29，30，32，33，36，37，39]reported the R0 resection rate， with 208 patients in the SMA-PD group and 170 in the S-PD group. There was a low heterogeneity between studies. In the fixed-effect model there was no difference between SMA-PD and S-PD (OR=1.43; 95% CI: 0.85 to 2.43; P>0.05). The R1 resection rate was reported by seven studies，[26，30，32，33，36，37，39]with low heterogeneity between them. There was no statistical significant difference between SMA-PD and S-PD (OR=0.85; 95% CI: 0.49 to 1.47; P>0.05).

We analyzed four studies[26，32，33，37]on pancreatic cancer only and also found no difference between SMA-PD and S-PD regarding the R0 (OR=1.56; 95% CI: 0.78 to 3.11; P>0.05) and R1 (OR=0.72; 95% CI: 0.36 to 1.44; P>0.05) resection rates; no differences between SMAPD and S-PD regarding one-year overall survival (OS) (OR=2.08; 95% CI: 0.57 to 7.52; P>0.05)， two-year OS (OR=0.93; 95% CI: 0.45 to 1.92; P>0.05)， and three-year OS (OR=1.24; 95% CI: 0.49 to 3.11; P>0.05).

The number of resected lymph nodes was reported by three studies，[27，30，32]with 59 patients in the SMA-PD group and 54 in the S-PD group. There was a high heterogeneity between the studies. The random-effects model revealed no statistical significant MD between SMA-PD and S-PD (MD=1.98; 95% CI: -2.41 to 6.37; P>0.05).

There was no difference regarding the extrapancreatic plexus invasion rate between SMA-PD and S-PD (OR=0.92; 95% CI: 0.46 to 1.85; P>0.05). This outcome was reported by four studies，[26，29，33，37]with low heterogeneity between them.

This outcome of increasing in circulating tumor cells was addressed only by Gall et al，[26]which proved a significant reduction of the circulating tumor cell by a notouch technique.

The one-year OS was reported by six studies，[26，29，32，33，36，37]with 142 patients in the SMA-PD group and 120 patients in the S-PD group. There was a moderate statistical heterogeneity between studies. In the random-effects model analysis there was no difference between SMA-PD and S-PD (OR=1.88; 95% CI: 0.83 to 4.23; P>0.05). The two-year OS was reported by five studies，[26，29，33，36，37]with 128 patients in the SMA-PD group and 95 patients in the S-PD group. There was a low heterogeneity betweenstudies. In the fixed-effect model analysis there was no difference between SMA-PD and S-PD (OR=1.19; 95% CI: 0.69 to 2.08; P>0.05). No difference was observed also for three-year OS (OR=0.89; 95% CI: 0.42 to 1.89; P>0.05). This outcome was reported by three studies，[29，33，37]with a low heterogeneity between them. There were 84 patients in the SMA-PD group and 71 in the S-PD group.

Fig. 3. Meta-analysis of studies on recurrences of patients undergoing superior mesenteric artery first approach pancreaticoduodenectomy (SMA-PD) versus standard pancreaticoduodenectomy (S-PD).

Two studies[26，39]reported the mean OS， which did not differed between SMA-PD and S-PD (MD=3.42; 95% CI: -0.26 to 7.10; P>0.05).

The local recurrence rate was reported by four studies，[32，33，36，37]with 111 patients in the SMA-PD group and 97 patents in the S-PD group (Fig. 3). There was a low heterogeneity between studies， with a significant reduction of the local recurrence rate in the SMA-PD group (OR=0.19; 95% CI: 0.08 to 0.44; P<0.01). Using Egger’s test， publication no bias was detected (t=0.57， P=0.31). The subgroup analysis of the three studies[32，33，37]which included only patients with pancreatic cancer showed a low heterogeneity between them. Fixed-effect model revealed the same decrease in local recurrence rate for SMA-PD (OR=0.14; 95% CI: 0.05 to 0.41; P<0.001). Using Egger’s test， publication no bias was detected (t=0.67， P=0.31).

The liver metastasis rate was reported by four studies，[26，32，33，36]with 77 patients in the SMA-PD group and 68 in the S-PD group. The statistical heterogeneity between studies was low. The SMA-PD was associated with a significant lower rate of liver metastasis (OR=0.43; 95% CI: 0.19 to 0.98; P=0.05). Using Egger’s test， no publication bias was detected (t=1.13， P=0.18). However， after Duval and Tweedie’s trim and fill analysis the 95% CI became narrower (OR=0.41; 95% CI: 0.20 to 0.86). The subgroup analysis for pancreatic cancer， reported by three studies，[26，32，33]showed no statistical significant difference for liver metastasis rate (OR=0.40; 95% CI: 0.15 to 1.08; P=0.07). For this subgroup also， no publication bias was detected (t=0.72， P=0.30).

The SMA-PD was associated with a lower rate of extrahepatic metastatic disease (OR=0.30; 95% CI: 0.10 to 0.93; P=0.04). This outcome was reported by three studies，[32，33，36]with low heterogeneity between them. There were 71 patients in the SMA-PD group and 62 in the S-PD group. Using Egger’s test， no publication bias was detected (t=1.43， P=0.19). Two studies[32，33]reported the extrahepatic metastatic rate in patients with pancreatic cancer with low heterogeneity between them (P=0.90， I2=0%). The SMA-PD was associated with a lower rate of extrahepatic metastatic disease (OR=0.22; 95% CI: 0.05 to 0.86; P=0.03). The publication bias can not be evaluated due to low number (only two) of pooled studies.

Rate of patients who received adjuvant chemotherapy， outcome which may associated with a better quality of life after surgery， was reported by five studies.[26，32，33，36，37]There were no differences between SMA-PD and S-PD (OR=2.19; 95% CI: 0.47 to 10.18; P>0.05).

Publication bias

Sensitivity analysis was conducted to assess statistical heterogeneity， through exclusion of specific studies with high risk of bias. There were no relevant changes in the overall effects of the quantitative synthesis. Analysis of the funnel plots did not show any significant asymmetries for any of the studied outcomes (Fig. 4).

Fig. 4. Begg’s funnel plot for the endpoint local recurrences (A)， liver metastasis (B)， and extrahepatic metastases (C) of patients undergoing superior mesenteric artery first approach pancreaticoduodenectomy (SMA-PD) versus standard pancreaticoduodenectomy (S-PD); blue circles-original studies， red circles-imputed filled studies using trim and fill method. Log odds ratio with 95% confidence interval before (blue diamond) and after (red diamond) allowing for publication bias.

Discussion

Our meta-analysis showed that SMA-PD for periampullary tumors is associated with better perioperative outcomes compared with S-PD， with a lower rate of local recurrences and metastatic spread. However， there was no benefits regarding one-， two- or three-year overall survival.

The SMA-PD was associated with the same rate of overall and major complications as the S-PD. We cannot explain the rate for lower pancreatic fistula rate in SMAPD， the proportion of pancreatic cancer patients being similar into the two groups (SMA-PD vs S-PD， 56.2% vs 51.8%). The diarrhea was significantly higher in the SMA-PD， especially in cases with left hemi-circumference SMA nervous plexus dissection. The technical feasibility of the SMA-PD was proved by the lack of differences regarding in-hospital mortality rate.

Identification and transection of the IPDA， usually located at 38±8.8 mm from the origin of the SMA， represents one of the key technical issues of a SMA first approach.[31，45]Analyzing imagistic data of 160 patients， Inoue et al[28]showed that IPDA emerged from a common trunk with the first jejunal artery in 71.6% of cases， independently from the SMA in 24.6% and from a replaced right hepatic artery in 3.8% of cases. The first jejunal vein run posterior to the SMA in 67.5%-83.3% of cases， two or more branches had an anterior and posterior course in 21.9% of cases and an anterior course in 10.6%-16.7% of patients.[28，46]

The current meta-analysis revealed that early inflow occlusion during pancreaticoduodenectomy is associated with significant reduction of intraoperative bleeding and transfusion requirements. This outcome has important benefits， intraoperative transfusion of one or two units of packed red blood cells being associated with increase in 30-day mortality， surgical-site infection， pneumonia， and sepsis in general surgery patients.[47]Red blood cell transfusion after pancreaticoduodenectomy has also important long-term oncological effects， being independently associated with earlier cancer recurrence and reduced survival， especially when administered postoperatively and in larger quantities.[48]

One of the major benefits of an SMA artery first approach seems to be accurate identification of the arterial anatomic variants， present in as many as 20% of cases， including a common hepatic artery from the SMA (2.34%) and a replaced right hepatic artery from the SMA (9.82%).[49，50]None of the included studies showed a higher rate for arterial iatrogenic injuries in S-PD compared with SMA-PD.

Starting from total mesorectal excision for rectal cancer[51]and complete mesocolic excision with central vascular ligation for colon cancer，[52]pancreatic surgeons proposed total mesopancreas excision for pancreatic cancer.[53]Total mesopancreas excision can be performed through a standard approach. Nevertheless， through a SMA first approach， the periadventiceal dissection guarantees a complete excision of the mesopancreas with central vascular ligation of the IPDA. The mesopancreas triangle， characterized by Adham and Singhirunnusorn， has a base lying on the posterior surface of the SMV and PV， a summit lying on the anterior surface of the aorta between celiac artery and SMA origin， and limited on each side by the right semi-circumferences of the celiac artery and SMA plexus.[17]Although it is not surrounded by a fibrous sheath or fascia， the mesopancreas should be accepted as a ‘’mesentery’’ of the head of the pancreas.[54]This connective tissue was defined by the Japan Pancreatic Society as pancreatic head nerve plexuses I and II.[55]

For patients with pancreatic cancer， the R0 resections are associated with a significant better median survival than R1 resections.[12]In the current meta-analysis the R0 resection rate was not significantly different between SMA-PD and S-PD. However， Kawabata et al[32]found that the R0 rate is higher in SMA-PD compared with that in S-PD. A report of Horiguchi et al published in Japanese showed also a higher R0 resection rate for SMAPD.[56]Using an anterior SMA first approach and a no touch isolation technique in patients with pancreatic cancer， Hirota et al[42]obtained a R0 resection rate of 82% and none R2， with a two-year survival of 75%. In 2014， the same group reported their experience of 52 patients with pancreatic cancer， operated on between 2008 and 2013， by a no-touch technique.[57]SMV-PV resections were necessary in twenty-one percent of patients with pancreatic head cancers. Analyzing all cases， there were 73% - R0， 21% - R1 and 6% - R2 resections. Fiveyear survival rate was 44% for pancreatoduodenectomy cases， significantly higher than the authors’ historical standard， of 13%.[57]

R0 resection is associated with dramatic survival benefit over R1 resection in a subset of patients with tumor size ≤25 mm and ≤1 involved lymphnodes.[58]Their median survival was 70.9 months compared with 17.7 months for R1 resections and 22.2 months for all the other R0 resections.[58]In the current study we did not find any statistical significant benefits regarding the one-， two-， or three-year OS for patients with SMA first approach. Maksymov et al[59]highlight the urgent need for standardized assessment of the pancreatoduodenectomy specimens. If they applied the College of the American Pathologists’ recommendations (assessment only of the bile duct， pancreatic neck and SMA margins)， an R1 status would be achieved in only 9 out of 25 patients. Extending the examination to the entire retroperitonealmargin (including the SMV margin and the posterior surface of the uncinate process margin) increased the rate of R1 resections to 14 out of 25 cases. Applying the 1-mm rule further increased the number of R1 to 20 of 25 patients.[59]A French multicenter prospective evaluation of resection margins in 150 specimens of pancreatoduodenectomy found 61% of cases being R1 resections， if the margin was defined as <1 mm.[12]The PV-SMV was the most frequently invaded resection margin， in 35% of cases. On multivariate analysis， the venous resection was the only factor correlated with a higher risk of at least one 0-mm positive resection margin.[12]Two-year progression free survival (PFS) and median PFS time in R0 and R1 (at 0 mm) groups were 42% vs 26.5%， and 19.5 vs 10.5 months， respectively (P=0.02). A positive PVSMV and SMA margin had significant impact on PFS， whereas a positive posterior margin had no impact.[12]Patients with a SMA-PD had a significantly higher rate of SMV-PV resection rate， which is actually one of the indications for such an approach.

Jang et al[60]randomized 200 patients to undergo standard or extended resection， the latter including dissection of additional lymph nodes and the right half of the nerve plexus around the SMA and celiac axis. The operation time and the estimated blood loss was higher in the extended resection group. The mean number of lymph nodes retrieved was higher in the extended group (33.7 vs 17.3， P<0.001) with a comparable R0 rate. The median survival after R0 resection was similar between the two groups (18 vs 19 months， P=0.239).[60]Our analysis found a similar number of retrieved lymphnodes and extrapancreatic plexus invasion rate. There are significant differences between different studies regarding the number of resected lymphnodes， and this can be explained by lack of standardization of pathology evaluation in some centers.[30]

An autopsy study， which analyzed the patterns of recurrence after curative resection of pancreatic cancer， showed that 75% of patients developed a local recurrence， 50% liver metastasis and 46% both.[61]Our study found that early dissection of the superior mesenteric artery is associated with a decrease not only of the local but also of the liver and extrahepatic metastatic rate. This may be due to the no-touch technique effect which was associated with decrease in circulating tumor cells[25]and carcinoembryonic antigen messenger RNA (mRNA)[62]in the portal vein.

Although was claimed as an important advantage to proceed to a SMA first approach， none of the included studies addressed the proportion of resections abandoned due to early detection of SMA invasion.

The current meta-analysis has several important limitations， especially due to clinical heterogeneity of the included studies， and caution should be exercised when interpreting its results. This meta-analysis involves several types of study designs， including retrospective， prospective， and randomized control trials， while the number of included patients is low. There was also an increased heterogeneity of the tumor type for which the PD was performed. Use of the random effects model for pooled data diminishes the effects of heterogeneity， and where observed fill and trim analysis tried to correct the publication biases.

In conclusion， SMA-PD for periampullary tumors is associated with better perioperative outcomes， such as blood loss， transfusion requirements， pancreatic fistula， and delayed gastric emptying without increasing duration of surgery， overall or major complication rate. Although the one-， two- or three-year overall survival rate is not superior， the superior mesenteric first approach pancreaticoduodenectomy has a lower local and metastatic recurrence rate.

Contributors: NI and BM proposed the study. HS and NRI performed the research and wrote the first draft. NI collected and analyzed the data. All authors contributed to the design and interpretation of the study and to further drafts. BM 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 Beuran M， Negoi I， Paun S， Ion AD， Bleotu C， Negoi RI， et al. The epithelial to mesenchymal transition in pancreatic cancer: A systematic review. Pancreatology 2015;15:217-225.

2 NCCN Clinical Practice Guidelines in Oncology. Pancreatic Adenocarcinoma. Pancreatic Adenocarcinoma. Available from: http://www.nccn.org/professionals/physician_gls/pdf/pancreatic.pdf. Published 2015 (Accessed July 14， 2015)

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

4 Orci LA， Meyer J， Combescure C， Bühler L， Berney T， Morel P， et al. A meta-analysis of extended versus standard lymphadenectomy in patients undergoing pancreatoduodenectomy for pancreatic adenocarcinoma. HPB (Oxford) 2015;17:565-572.

5 Fortner JG. Regional pancreatectomy for cancer of the pancreas， ampulla， and other related sites. Tumor staging and results. Ann Surg 1984;199:418-425.

6 Pessaux P， Regenet N， Arnaud JP. Resection of the retroportal pancreatic lamina during a cephalic pancreaticoduodenectomy: first dissection of the superior mesenteric artery. Ann Chir 2003;128:633-636.

7 Jimenez R， Warshaw A. Pancreaticoduodenectomy for pancreatic cancer: results after Kausch-Whipple and pyloruspreserving resection. In: Beger H， Warshaw A， Büchler M， etal.， eds. The Pancreas: An Integrated Textbook of Basic Science， Medicine， and Surgery. 2nd ed. Oxford， UK: Blackwell Publishing Limited; 2008:696-706.

8 Pessaux P， Varma D， Arnaud JP. Pancreaticoduodenectomy: superior mesenteric artery first approach. J Gastrointest Surg 2006;10:607-611.

9 Al-Hawary MM， Francis IR， Chari ST， Fishman EK， Hough DM， Lu DS， et al. Pancreatic ductal adenocarcinoma radiology reporting template: consensus statement of the Society of Abdominal Radiology and the American Pancreatic Association. Radiology 2014;270:248-260.

10 Valls C， Andía E， Sanchez A， Fabregat J， Pozuelo O， Quintero JC， et al. Dual-phase helical CT of pancreatic adenocarcinoma: assessment of resectability before surgery. AJR Am J Roentgenol 2002;178:821-826.

11 Sanjay P， Takaori K， Govil S， Shrikhande SV， Windsor JA.‘Artery-first’ approaches to pancreatoduodenectomy. Br J Surg 2012;99:1027-1035.

12 Delpero JR， Bachellier P， Regenet N， Le Treut YP， Paye F， Carrere 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.

13 Verbeke CS， Menon KV. Redefining resection margin status in pancreatic cancer. HPB (Oxford) 2009;11:282-289.

14 Esposito I， Kleeff J， Bergmann F， Reiser C， Herpel E， Friess H， et al. Most pancreatic cancer resections are R1 resections. Ann Surg Oncol 2008;15:1651-1660.

15 Gaedcke J， Gunawan B， Grade M， Sz?ke R， Liersch T， Becker H， et al. The mesopancreas is the primary site for R1 resection in pancreatic head cancer: relevance for clinical trials. Langenbecks Arch Surg 2010;395:451-458.

16 Sohn TA， Yeo CJ， Cameron JL， Koniaris L， Kaushal S， Abrams RA， et al. Resected adenocarcinoma of the pancreas-616 patients: results， outcomes， and prognostic indicators. J Gastrointest Surg 2000;4:567-579.

17 Adham M， Singhirunnusorn J. Surgical technique and results of total mesopancreas excision (TMpE) in pancreatic tumors. Eur J Surg Oncol 2012;38:340-345.

18 Rose JB， Rocha F， Alseidi A， Helton S. Posterior ‘superior mesenteric artery first’ approach for resection of locally advanced pancreatic cancer. Ann Surg Oncol 2014;21:1927-1928.

19 Liberati A， Altman DG， Tetzlaff J， Mulrow C， G?tzsche PC， Ioannidis JP， et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ 2009;339:b2700.

20 The Cochrane Collaboration. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [Updated March 2011]. Available from: www.cochrane-handbook.org (Accessed July 11， 2015)

21 Sterne J， Higgins J， Reeves B， on behalf of the development group for ACROBAT-NRSI. A Cochrane Risk Of Bias Assessment Tool: for Non-Randomized Studies of Interventions (ACROBAT-NRSI). Version 1.0.0. Available from: http://www. riskofbias.info. (Accessed July 15， 2015)

22 Hozo SP， Djulbegovic B， Hozo I. Estimating the mean and variance from the median， range， and the size of a sample. BMC Med Res Methodol 2005;5:13.

23 Chootrakool H， Shi JQ， Yue R. Meta-analysis and sensitivity analysis for multi-arm trials with selection bias. Stat Med 2011;30:1183-1198.

24 DerSimonian R， Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177-188.

25 Egger M， Davey Smith G， Schneider M， Minder C. Bias in meta-analysis detected by a simple， graphical test. BMJ 1997;315:629-634.

26 Gall TM， Jacob J， Frampton AE， Krell J， Kyriakides C， Castellano L， et al. Reduced dissemination of circulating tumor cells with no-touch isolation surgical technique in patients with pancreatic cancer. JAMA Surg 2014;149:482-485.

27 Zhou HY， Wang Y， Zhang J， Ruan CP， Wang WJ， Sun YP， et al. Retrograde vs conventional dissection technique in pancreaticoduodenectomy: a pilot study. JAMA Surg 2014;149:604-607.

28 Inoue Y， Saiura A， Yoshioka R， Ono Y， Takahashi M， Arita J， et al. Pancreatoduodenectomy with systematic mesopancreas dissection using a supracolic anterior artery-first approach. Ann Surg 2015;262:1092-1101.

29 Gundara JS， Wang F， Alvarado-Bachmann R， Williams N， Choi J， Gananadha S， et al. The clinical impact of early complete pancreatic head devascularisation during pancreatoduodenectomy. Am J Surg 2013;206:518-525.

30 Shrikhande SV， Barreto SG， Bodhankar YD， Suradkar K， Shetty G， Hawaldar R， et al. Superior mesenteric artery first combined with uncinate process approach versus uncinate process first approach in pancreatoduodenectomy: a comparative study evaluating perioperative outcomes. Langenbecks Arch Surg 2011;396:1205-1212.

31 Kawai M， Tani M， Ina S， Hirono S， Nishioka R， Miyazawa M， et al. CLIP method (preoperative CT image-assessed ligation of inferior pancreaticoduodenal artery) reduces intraoperative bleeding during pancreaticoduodenectomy. World J Surg 2008;32:82-87.

32 Kawabata Y， Tanaka T， Nishi T， Monma H， Yano S， Tajima Y. Appraisal of a total meso-pancreatoduodenum excision with pancreaticoduodenectomy for pancreatic head carcinoma. Eur J Surg Oncol 2012;38:574-579.

33 Aimoto T， Mizutani S， Kawano Y， Matsushita A， Yamashita N， Suzuki H， et al. Left posterior approach pancreaticoduodenectomy with total mesopancreas excision and circumferential lymphadenectomy around the superior mesenteric artery for pancreatic head carcinoma. J Nippon Med Sch 2013;80:438-445.

34 Shah OJ， Gagloo MA， Khan IJ， Ahmad R， Bano S. Pancreaticoduodenectomy: a comparison of superior approach with classical Whipple’s technique. Hepatobiliary Pancreat Dis Int 2013;12:196-203.

35 Ishizaki Y， Sugo H， Yoshimoto J， Imamura H， Kawasaki S. Pancreatoduodenectomy with or without early ligation of the inferior pancreatoduodenal artery: comparison of intraoperative blood loss and short-term outcome. World J Surg 2010;34:2939-2944.

36 Figueras J， Codina-Barreras A， López-Ben S， Maroto A， Torres-Bahí S， González HD， et al. Cephalic duodenopancreatectomy in periampullary tumours. Dissection of the superior mesenteric artery as aninitial approach. Description of the technique and an assessment of our initial experience. Cir Esp 2008;83:186-193.

37 Kurosaki I， Minagawa M， Takano K， Takizawa K， Hatakeyama K. Left posterior approach to the superior mesenteric vascular pedicle in pancreaticoduodenectomy for cancer of the pancreatic head. JOP 2011;12:220-229.

38 Horiguchi A， Ishihara S， Ito M， Nagata H， Shimizu T， Furusawa K， et al. Pancreatoduodenectomy in which dissection of the efferent arteries of the head of the pancreas is performed first. J Hepatobiliary Pancreat Surg 2007;14:575-578.

39 Dumitrascu T， David L， Popescu I. Posterior versus standard approach in pancreatoduodenectomy: a case-match study. Langenbecks Arch Surg 2010;395:677-684.

40 Marzano E， Piardi T， Marescaux J， Pessaux P. Combined posterior and anterior approach to the superior mesenteric artery: the advantages of the “hanging maneuver”. Langenbecks Arch Surg 2012;397:1023-1024.

41 Pessaux P， Rosso E， Panaro F， Marzano E， Oussoultzoglou E， Bachellier P， et al. Preliminary experience with the hanging maneuver for pancreaticoduodenectomy. Eur J Surg Oncol 2009;35:1006-1010.

42 Hirota M， Kanemitsu K， Takamori H， Chikamoto A， Tanaka H， Sugita H， et al. Pancreatoduodenectomy using a no-touch isolation technique. Am J Surg 2010;199:e65-68.

43 Weitz J， Rahbari N， Koch M， Büchler MW. The “artery first”approach for resection of pancreatic head cancer. J Am Coll Surg 2010;210:e1-4.

44 Hackert T， Werner J， Weitz J， Schmidt J， Büchler MW. Uncinate process first--a novel approach for pancreatic head resection. Langenbecks Arch Surg 2010;395:1161-1164.

45 Cho A， Yamamoto H， Kainuma O. Tips of laparoscopic pancreaticoduodenectomy: superior mesenteric artery first approach (with video). J Hepatobiliary Pancreat Sci 2014;21:E19-21.

46 Nakamura M， Nakashima H， Tsutsumi K， Matsumoto H， Muta Y， Ueno D， et al. First jejunal vein oriented mesenteric excision for pancreatoduodenectomy. J Gastroenterol 2013;48:989-995.

47 Bernard AC， Davenport DL， Chang PK， Vaughan TB， Zwischenberger JB. Intraoperative transfusion of 1 U to 2 U packed red blood cells is associated with increased 30-day mortality， surgical-site infection， pneumonia， and sepsis in general surgery patients. J Am Coll Surg 2009;208:931-939.

48 Kneuertz PJ， Patel SH， Chu CK， Maithel SK， Sarmiento JM， Delman KA， et al. Effects of perioperative red blood cell transfusion on disease recurrence and survival after pancreaticoduodenectomy for ductal adenocarcinoma. Ann Surg Oncol 2011;18:1327-1334.

49 Yang SH， Yin YH， Jang JY， Lee SE， Chung JW， Suh KS， et al. Assessment of hepatic arterial anatomy in keeping with preservation of the vasculature while performing pancreatoduodenectomy: an opinion. World J Surg 2007;31:2384-2391.

50 Pallisera A， Morales R， Ramia JM. Tricks and tips in pancreatoduodenectomy. World J Gastrointest Oncol 2014;6:344-350.

51 How P， Shihab O， Tekkis P， Brown G， Quirke P， Heald R， et al. A systematic review of cancer related patient outcomes after anterior resection and abdominoperineal excision for rectal cancer in the total mesorectal excision era. Surg Oncol 2011;20: e149-155.

52 West NP， Kobayashi H， Takahashi K， Perrakis A， Weber K， Hohenberger W， et al. Understanding optimal colonic cancer surgery: comparison of Japanese D3 resection and European complete mesocolic excision with central vascular ligation. J Clin Oncol 2012;30:1763-1769.

53 Chowdappa R， Challa VR. Mesopancreas in pancreatic cancer: where do we stand - review of literature. Indian J Surg Oncol 2015;6:69-74.

54 Agrawal MK， Thakur DS， Somashekar U， Chandrakar SK， Sharma D. Mesopancreas: myth or reality? JOP 2010;11:230-233.

55 Mizuno S， Isaji S， Tanemura A， Kishiwada M， Murata Y， Azumi Y， et al. Anterior approach to the superior mesenteric artery by using nerve plexus hanging maneuver for borderline resectable pancreatic head carcinoma. J Gastrointest Surg 2014;18:1209-1215.

56 Horiguchi A， Ishihara S， Ito M， Asano Y， Yamamoto T， Miyakawa S. Pancreatoduodenectomy for invasive carcinoma of the head of the pancreas with ligation of the inferior pancreaticoduodenal artery performed first. Nihon Geka Gakkai Zasshi 2011;112:159-163.

57 Hirota M， Ogawa M. No-touch pancreatectomy for invasive ductal carcinoma of the pancreas. JOP 2014;15:243-249.

58 Tummala P， Howard T， Agarwal B. Dramatic survival benefit related to R0 resection of pancreatic adenocarcinoma in patients with tumor ≤25 mm in size and ≤1 involved lymph nodes. Clin Transl Gastroenterol 2013;4:e33.

59 Maksymov V， Hogan M， Khalifa MA. An anatomical-based mapping analysis of the pancreaticoduodenectomy retroperitoneal margin highlights the urgent need for standardized assessment. HPB (Oxford) 2013;15:218-223.

60 Jang JY， Kang MJ， Heo JS， Choi SH， Choi DW， Park SJ， et al. A prospective randomized controlled study comparing outcomes of standard resection and extended resection， including dissection of the nerve plexus and various lymph nodes， in patients with pancreatic head cancer. Ann Surg 2014;259:656-664.

61 Hishinuma S， Ogata Y， Tomikawa M， Ozawa I， Hirabayashi K， Igarashi S. Patterns of recurrence after curative resection of pancreatic cancer， based on autopsy findings. J Gastrointest Surg 2006;10:511-518.

62 Hirota M， Shimada S， Yamamoto K， Tanaka E， Sugita H， Egami H， et al. Pancreatectomy using the no-touch isolation technique followed by extensive intraoperative peritoneal lavage to prevent cancer cell dissemination: a pilot study. JOP 2005;6:143-151.

Received December 31， 2015

Accepted after revision July 22， 2016

Author Affiliations: General Surgery Department， Emergency Hospital of Bucharest (Negoi I， Runcanu A and Beuran M); National Institute of Legal Medicine Mina Minovici (Hostiuc S)， and Anatomy Department (Negoi RI)， Carol Davila University of Medicine and Pharmacy Bucharest， Bucharest， Romania

Ionut Negoi， MD， PhD， Senior Lecturer of Surgery， General Surgery Department， Emergency Hospital of Bucharest， Carol Davila University of Medicine and Pharmacy Bucharest， No 8 Floreasca Street， Sector 1， 014461， Bucharest， Romania (Tel: +40723209910; Fax: +40215992257; Email: negoiionut@gmail.com)

? 2017， Hepatobiliary Pancreat Dis Int. All rights reserved.

10.1016/S1499-3872(16)60134-0

Published online September 13， 2016.