minimally-invasive vs open pancreaticoduodenectomy...

Minimally-Invasive vs OpenPancreaticoduodenectomy: Systematic Reviewand Meta-Analysis

Camilo Correa-Gallego, MD, Helen E Dinkelspiel, MD, Isabel Sulimanoff, MLS, Sarah Fisher, MD,Eduardo F Vinuela, MD, T Peter Kingham, MD, FACS, Yuman Fong, MD, FACS,Ronald P DeMatteo, MD, FACS, Michael I D’Angelica, MD, William R Jarnagin, MD, FACS,Peter J Allen, MD, FACS

Laparoscopic approaches are routinely used for a variety ofprocedures in general surgery and various surgical special-ties including surgical oncology. Since publication of thefirst series of laparoscopic cholecystectomy in the late1980s, the field of minimally invasive surgery (MIS) hasexpanded dramatically and is now regarded as an estab-lished specialty. Many oncologic procedures have provednot only feasible and safe, but oncologically equivalentto traditional open procedures regarding both immediateoperative variables of interest (margins, lymph noderetrieval, and morbidity) and long-term outcomes.1-10

Pancreaticoduodenectomy (PD) poses a particularchallenge. During this procedure, there is extensive retro-peritoneal dissection around anatomically complex andhazardous structures, and a prolonged reconstructionthat includes 3 technically demanding anastomoses.Given this complex gastrointestinal reconstruction, ithas been generally thought that the minimally invasiveapproach would not significantly decrease recovery time(hospital stay), yet it would significantly increase opera-tive time. Even though minimally invasive PD wasreported as early as 1994,11 laparoscopic surgeons havebeen reluctant to routinely perform it. Since this firstdescription now almost 20 years ago, a large number ofsingle-institution series of minimally invasive (includinglaparoscopic-assisted, totally laparoscopic, and more

recently robotic) PD performed for a variety of indica-tions have been reported.12-27

Based on safety data derived from those retrospectivereports, these procedures are now offered to selectedpatients at a limited number of institutions.However, thereis currently no level 1 evidence that compares outcomesbetween MIS and the traditional open approach. Wepresent a systematic review of the literature and a compara-tive effectiveness analysis ofminimally invasive vs open PD.Using meta-analytic techniques, we set out to evaluate thesurgical and oncologic outcomes of patients undergoingeither procedure as reported in the published literature.

METHODSThis study adheres to the Preferred Reporting Items forSystematic Reviews and Meta-Analyses (PRISMA) guide-lines.28 We queried medical databases in search for manu-scripts comparing operative, postoperative, and oncologicoutcomes of minimally invasive and open PD. An experi-enced information scientist (IS) designed the searchstrategy, which was refined and revised by a surgeon (CC).

Literature search

Comprehensive literature searches were performed of thePubMed, EMBASE, and Cochrane Library databases forthe years 1994 through January 2013. No language restric-tion was imposed. Three categories of terms were“ANDED” together: pancreaticoduodenectomy terms,laparoscopy/robot-assisted/minimally invasive terms, andoutcomes terms. For PubMed, a search using MedicalSubject Headings (MeSH) terms was run, as well as a text-word search. For EMBASE, a search using Emtree(EMBASE vocabulary) terms was run, as was a textwordsearch. For Cochrane, a search using MeSH terms wasrun, as well as a textword search. All retrieved recordswere added to an EndNote (Version X6 e Thomson Reu-ters) library. The last search was performed on January 15,2013. A total of 703 references were retrieved, and afterremoval of duplicates, 527 references remained.

CME questions for this article available athttp://jacscme.facs.org

Disclosure Information: Authors have nothing to disclose. Timothy J Eberlein,Editor-in-Chief, has nothing to disclose.

Received July 17, 2013; Revised August 27, 2013; Accepted September 16,2013.From the Department of Surgery (Correa-Gallego, Vinuela, Kingham,Fong, DeMatteo, D’Angelica, Jarnagin, Allen) and the Medical Library(Sulimanoff), Memorial Sloan-Kettering Cancer Center; and the Divisionof Gynecologic Oncology (Dinkelspiel), New York Presbyterian Hospital,Weill Cornell Medical Center, New York, NY; and the Department ofSurgery (Fisher), Emory University Hospital, Atlanta, GA.Correspondence address: Peter J Allen, MD, FACS, 1275 York Ave,Memorial Sloan-Kettering Cancer Center, New York, NY. email: [email protected]

129ª 2014 by the American College of Surgeons ISSN 1072-7515/13/$36.00

Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jamcollsurg.2013.09.005

mailto:[email protected]

mailto:[email protected]

http://dx.doi.org/10.1016/j.jamcollsurg.2013.09.005

Study selection

Minimally invasive PD was defined as completely laparo-scopic or robotic resection of the head of the pancreas andduodenum, followed by completely intracorporeal recon-struction of the pancreatic, biliary, and intestinal conti-nuity. Hybrid procedures, in which part of thedissection or reconstruction was extracorporeal orthrough a “mini-laparotomy” were not considered forthis study.Studies were included only if they were original series

comparing minimally invasive and open PD thatreported at least 1 of the outcomes variables of interest.We included only published studies dating back to 1994(first report of minimally invasive PD) regardless oflanguage.Studies were excluded from our analysis if they did not

report at least 1 of the outcomes of interest, includedminimally invasive or open PDs but not both, or didnot perform a comparison between the 2 techniques;reported hybrid procedures; reported on fewer than 8cases; if they were technical “how-to” reports, or if theywere animal or unpublished studies for which completedata for pooling were not available.Two authors (CC and HD) evaluated all titles to iden-

tify relevant articles. Abstracts of these were re-evaluatedto identify those meeting inclusion criteria, and full textsof these articles were obtained for data extraction.Disagreement was solved by a third author (SF).

Data extraction

Two authors independently extracted the data of interestfrom the manuscripts. Results were compared andconsensus was reached. The outcomes of interest forour study were: operative/pathology variables (operativetime, estimated blood loss, R0 resection, lymph nodeharvest, and tumor size), postoperative outcomes (overallcomplications, pancreatic fistula rate [and grade �3],delayed gastric emptying, postpancreatectomy hemor-rhage, wound infection, length of hospital stay, and reop-eration rate). Operative outcomes largely determine theultimate oncologic outcomes of these patients (in thesetting of malignancy); the postoperative outcomes evalu-ated represent the major drivers of morbidity after pan-creaticoduodenectomy. Because our search was notlimited to a specific indication for this operation,long-term oncologic outcomes were not included in our

analysis. The reference sections of the selected studieswere searched for additional relevant papers. Correspond-ing authors were not contacted.

Statistical analysis

The meta-analysis was conducted according to the recom-mendations of the Preferred Reporting Items for System-atic Reviews and Meta-Analysis (PRISMA) Statement.28

Whenever missing from a study, mean and standard devi-ation were estimated from the data available, if possible,as previously described.29 Studies were combined usinga random-effects model and estimates were expressed asweighted mean differences (WMD) for continuous dataand odds ratios (OR) for event-related outcomes. In addi-tion, 95% confidence intervals (CI) were calculated. I2

values were preferred for quantification of statisticalinconsistency, defined as the percentage of variationbetween studies due to heterogeneity.30 Analyses were per-formed with Comprehensive MetaAnalysis Version 2.0(Biostat Solutions Inc).

Assessment of methodologic quality of ourmeta-analytic techniques

The previously validated Overview Quality AssessmentQuestionnaire (OQAQ)31 was used to assess the meth-odologic rigor of our study. This questionnaire includesa self-reported subset of questions regarding items thatshould be fulfilled to ensure quality: 1. Was the searchcomprehensive? 2. Was selection bias avoided? 3. Wasvalidity assessed appropriately? and 4. Were the methodsused to combine results from studies appropriate?

RESULTSThe predefined inclusion criteria were met by 6 studiesthat included 542 patients (169 MIS and 373 open)and these were pooled in the meta-analysis.32-37 APRISMA flow diagram depicting the selection process isshown in Figure 1. General study characteristics aredetailed in Table 1. There were no randomized controlledtrials identified. All the studies found were retrospectivereviews of variable quality that compared consecutivecases of minimally invasive PD with either consecutiveor matched open procedures performed during thesame time period. All indications were included; however,the majority of cases were performed for malignancy. Allstudies focused on operative and perioperative outcomes,there were no reports of long-term oncologic results, andthere were no multicenter studies.

Meta-analysis

Results of the meta-analysis are summarized in Table 2.Forest plots of those comparisons that showed a significant

Abbreviations and Acronyms

MIS ¼ minimally invasive surgeryPD ¼ pancreaticoduodenectomyWMD ¼ weighted mean difference

130 Correa-Gallego et al Minimally Invasive vs Open Pancreaticoduodenectomy J Am Coll Surg

difference are shown in Figures 2 to 7. Minimally invasivesurgery was associated with a reduction in intraoperativeblood loss (WMD 1,460 mL, 95% CI 726 to 2,194 mL,p < 0.001, I2 ¼ 95%), although longer operative timeswere observed (WMD 131 minutes, 95% CI 43 to

218 minutes, p ¼ 0.003, I2 ¼ 96%). In addition, inthe MIS group, retrieval of lymph nodes was significantlyhigher by 3 nodes (95% CI 0.3 to 6.0 nodes, p ¼ 0.03,I2 ¼ 61%) and the likelihood of an R1 resection inpatients with malignancy was lower (OR 0.4, 95%

Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram.

Table 1. General Study Characteristics

Indication, n

Approach, n Benign Malignant

First author Year Study design Total n MIS Open MIS Open MIS Open

Asbun32 2012 Retrospective 268 53 215 14 74 39 141

Buchs33 2011 Retrospective 83 44 39 11 12 33 27

Chalikonda34 2012 Retrospective 60 30 30 16 16 14 14

Lai35 2012 Retrospective 87 20 67 5 14 15 53

Zhou36 2011 Retrospective 16 8 8 0 2 8 6

Zureikat37 2011 Retrospective 28 14 14 2 2 12 12

MIS, minimally invasive surgery.

print&web4C/FPO

Vol. 218, No. 1, January 2014 Correa-Gallego et al Minimally Invasive vs Open Pancreaticoduodenectomy 131

CI 0.2 to 0.8, p ¼ 0.007, I2 ¼ 0%). Of note, tumor sizewas significantly larger in the open group (WMD 0.6 cm,95% CI 0.1 to 1.1 cm, p ¼ 0.02, I2 ¼ 51%). Thirty-fiveof 169 patients (21%) in the MIS group and 63 of 373(17%) in the open group developed a pancreatic fistula;

for clinically relevant pancreatic fistula, the proportionswere 13 of 169 (8%) and 26 of 373 (7%), respectively;p ¼ 0.9 (Table 2 and Fig. 8). Overall morbidity, reopera-tions, and other specific complications (delayed gastricemptying, postpancreatectomy hemorrhage, and wound

Table 2. Meta-Analysis of Different Outcomes Variables

Variable Studies includedPatients*

OR/WMD 95% CI p Value I2, %MIS/Open

Operative time, min 6 169/373 131 43e218 0.003y 96

Blood loss, mL 6 169/373 (�) 1,460 (�) 2,194e(�) 726 <0.001y 95

Harvested lymph nodes 5 161/365 (�) 3.2 (�) 6.0e(�) 0.3 0.03y 61

R1/R2 resectionz 6 121/253 0.4 0.20e0.80 0.007y 0

Tumor size, cm 4 117/326 0.6 0.1e1.1 0.02y 51

Overall complications 6 169/373 0.67 0.39e1.16 0.15 39

PF Overall 6 169/373 1.11 0.68e1.83 0.67 0

PF Clinically relevant 6 169/373 0.97 0.48e1.99 0.94 0

DGE 4 147/351 0.75 0.35e1.63 0.48 0

PPH 4 147/351 1.55 0.70e3.42 0.28 0

Wound infection 4 147/351 0.49 0.23e1.1 0.077 0

Reoperation 5 155/359 0.54 0.25e1.18 0.12 0

Hospital stay 5 139/343 (�) 3.7 (�) 6.8e(�) 0.5 0.02y 78

*Total number of patients evaluated in each category.ySignificant statistical difference.zCalculated over the number of patients with a malignancy.DGE, delayed gastric empting; MIS, minimally invasive surgery; OR, odds ratio; PF, pancreatic fistula; PPH, post-pancreatectomy hemorrhage; WMD,weighted mean difference.

Figure 2. Pooled estimate of operative time for minimally invasive surgery (MIS) vs openpancreaticoduodenectomy. All studies except for Buch’s show a statistically significant longeroperative time for MIS procedures.

print&web4C/FPO


infection) were also comparable (Table 2). Hospital staywas significantly reduced in the MIS group by 3.7 days(95% CI 0.5 to 6.8 days, p ¼ 0.02, I2 ¼ 78%).A total of 17 patients in the MIS group (10%) required

conversion to an open procedure. The majority of these(13 of 17) were due to bleeding or the inability to obtainadequate vascular control laparoscopically; there was 1perioperative death reported that was directly related to

surgical hemorrhage. Only 2 of the studies includedpatients undergoing conversion in the MIS arm in anintention-to-treat analysis.

Assessment of methodologic quality and bias

All the studies identified were case series. Because therehave been no randomized controlled trials of MIS vsopen PD, the Cochrane Collaboration’s tool for assessing

Figure 3. Pooled estimate of estimated operative blood loss for minimally invasive surgery(MIS) vs open pancreaticoduodenectomy.

Figure 4. Pooled estimate of number of harvested lymph nodes during minimally invasivesurgery (MIS) vs open pancreaticoduodenectomy.

print&web4C/FPO

print&web4C/FPO


risk of bias38 was not applied. The studies included in ourmeta-analysis represent the experience of single referralcenters and in most cases, a single surgeon’s experience.Furthermore, they correspond to the initial experiencewith this complex minimally invasive technique and donot account for the learning curve to perform these proce-dures. In most cases, the selection criteria for patients toundergo MIS were not reported.

DISCUSSIONContinuous improvement and technical advances (inequipment as well as training) have made possible theapplication of minimally invasive techniques to a widevariety of surgical procedures, including complex onco-logic operations.1-10 Properly conducted randomized trialshave shown noninferiority of MIS compared with openapproaches, as long as oncologic principles are preserved.The identified benefits of minimally invasive oncologicresection have been a smaller incision and shorterrecovery time. However, to date there have been no iden-tified benefits associated with these less invasive proce-dures with regard to oncologic outcomes.In surgical oncology, minimally invasive approaches do

not (and should not) significantly modify the operativetechniques or the steps defining the operation itself; the

only significant difference linked to MIS lies in the waythese goals are achieved. The expectation that a differentincision would significantly improve the outcomes of anotherwise identical operation is therefore not realistic.Although early wound complications (infections, dehis-cence, etc) are a major source of distress for generalsurgical patients, the 2 major drivers of morbidity afterPD are failure of the pancreatic anastomosis leading topancreatic fistulae and delayed gastric emptying. Bothof these complications are unrelated to the abdominalincision. On the other hand, late wound complications(eg, incisional hernias) have been identified as a sourceof delayed morbidity after pancreatic resection39;however, the morbidity associated with an incisionalhernia must be interpreted in the context of pancreaticmalignancy.Minimally invasive surgery techniques have shown to

be associated with lower wound infection rates indifferent kinds of abdominal operations, and this is oneof its main advantages.40,41 Four of the studies includedin our meta-analysis report specifically on wound infec-tion rates. We found a nonsignificant trend toward fewerwound infections in patients who underwent minimallyinvasive PD (OR 0.49, 95% CI 0.23 to 1.1,p ¼ 0.077, I2 ¼ 0%). Theoretically, decreased woundcomplication rates (and decreased morbidity in general)

Figure 5. Pooled estimate of rate of R0 resections for minimally invasive surgery (MIS) vs openpancreaticoduodenectomy.

print&web4C/FPO


have the potential to allow eligible patients to initiatepostoperative chemotherapy in a timely manner, andtherefore possibly contribute to an improved oncologicoutcome. However, these theoretical advantages areunproven, and the studies included in our pooled analysis

did not address this issue specifically because the indica-tions for resection (and the need for adjuvant therapy)were variable and the long-term oncologic outcomeswere not reported. Furthermore, despite a trend towardimproved wound infection rates, there was no difference

Figure 7. Pooled estimate of length of hospital stay (days) for minimally invasive surgery (MIS)vs open pancreaticoduodenectomy.

Figure 6. Pooled estimate of tumor size (cm) for minimally invasive surgery (MIS) vs openpancreaticoduodenectomy.

print&web4C/FPO

print&web4C/FPO


Figure 8. (A) Pooled estimate of incidence of postoperative pancreatic fistula and (B) clinically-relevant postoperative pancreatic fistula for minimally invasive surgery (MIS) vs openpancreaticoduodenectomy.

print&

web4C/FPO


in overall complications (Table 2).These important (and yetunanswered) questions underscore the need for adequatelydesigned trials that evaluate the MIS approach in patientswith prespecified inclusion and matching criteria.The findings in this study indicate that minimally inva-

sive PD is a feasible procedure in highly selected patientswho are cared for in select centers. These data suggest thatminimally invasive PD may be associated with a decreasedhospital stay, and possibly higher lymph node retrieval.The findings of lower operative blood loss and higherrate of R0 resection must be viewed in the setting ofhighly selected patients with significantly smaller tumors.The likelihood of bias is significant, and not untilrandomized data are available, or retrospective data thatdemonstrate clear outcome equivalence or improvement,can this approach be considered standard.Operative blood loss was shown in the pooled analysis

to be lower in MIS cases. This is often attributed to themagnified view of small vessels that laparoscopy allowsfor, particularly during dissection of the plane betweenthe uncinate process and the superior mesenteric vessels.It is important to underscore though, that all the studiesin this meta-analysis systematically excluded patients withknown vascular involvement for consideration of the MISapproach, and those with unexpected vascular involve-ment were converted to an open approach. Lack ofadequate matching in this regard makes comparison ofoperative blood loss inherently flawed and at a high riskfor confounding. Furthermore, not all studies performedan intention-to-treat analysis and included patients whorequired conversion from MIS to an open approach inthe open arm for the analysis. This introduces consider-able bias because these patients have expected higherblood loss and longer operating times. Dissection of theporto-mesenteric axis is an intrinsically challenging andhazardous part of PD, and failure to expeditiously controlthese vessels may lead to massive bleeding and exsangui-nation. Therefore, it is of the utmost importance thatsurgeons embarking on this operation are capable ofcontrolling hemorrhage laparoscopically and that theresources are readily available for conversion toa laparotomy.The median postoperative length of stay of patients

undergoing minimally invasive PD was 11 days (inter-quartile range: 8 to 14 days); it was 14 days (interquartilerange: 12 to 24 days) for open cases. Most current seriesfrom referral centers report median stays of 7 or 8 daysafter open PD, and the considerably longer postoperativestays may have been related to local preferences anddifferent health care systems.42-44 Furthermore, the shorterrecovery period in these patients did not seem to translateinto (or be explained by) lower postoperative morbidity.

The current median length of stay at our institution afteropen PD is 8 days,45 which is 3 days shorter than that inthe MIS groups reported by the studies included in ourmeta-analysis.Although long-term oncologic outcomes are not

addressed in these early studies, R0 status and lymphnode retrieval are used as indicators of the oncologicadequacy of MIS techniques. Lymph node staging isa very relevant prognostic factor for pancreatic cancerpatients. However, different pathologic processing tech-niques of the surgical specimen have been shown to yieldsignificantly different lymph node counts.46 Without anindication of how the specimens were processed in thesestudies, it’s hard to draw strong conclusions in thisregard. Moreover, the higher R0 resection rate seenshould be interpreted with caution. Our pooled analysisalso showed that patients in the MIS group had smallertumors, which must be considered a confounder thatmay account for the higher R0 resection rates.To date there are no published randomized controlled

trials comparing minimally invasive (either robotic orplain laparoscopic) PD, nor are there, to our knowledge,any registered trials ongoing. Enthusiasm has beenprimarily guided by a small number of surgeons at largevolume centers. There is no doubt that minimally inva-sive techniques have earned a significant role in the treat-ment of abdominal malignancies and there is no reason tobelieve that the indications will not continue to expandinto PD as instrumentation and training continue toevolve.47 However, it is at this time of equipoise thata well-designed prospective study of minimally invasivePD could be accomplished at select centers. The valueof quality clinical trial research will become even moreimportant in the upcoming era of cost containment,when our interventions are likely to be subject to muchtighter scrutiny by regulatory agencies.48

Our study has several limitations. First, there are norandomized controlled trials, and no prospective studiesof high quality that provide unbiased data for our anal-ysis. This is, however, not only a limitation of our study,but an important indication that the widespread applica-tion of these techniques is yet to be achieved. The firstlaparoscopic PD was reported almost 20 years ago, andthe lack of penetrance of this approach into the vastmajority of institutions should be given consideration.Additionally, there is high variability between the studieswe analyzed, which is addressed (albeit incompletely) bythe random-effects model we used for our meta-analysis. Furthermore, given the paucity of data, wegrouped together the outcomes of pure laparoscopicand robotic-assisted cases, which may be intrinsicallydifferent between themselves. Another major limitation


lies in the possibility of publication bias, in which centersand individual surgeons who have had positive outcomeswith MIS major pancreatic resection are more likely topublish their findings. These limitations notwithstanding,we believe that our study fulfills the Overview QualityAssessment Questionnaire evaluation and providesimportant information regarding the very limited avail-able data on the outcomes of minimally invasive PD.

CONCLUSIONSMinimally invasive PD is currently a feasible operation inselect patients being treated at select centers. Although thefindings of this meta-analysis suggest there might beimproved outcomes associated with minimally invasivePD in comparison with the open approach, the lack ofrandomized trials or high-quality, nonrandomizedprospective studies comparing both techniques does notallow for firm conclusions to be drawn, so minimallyinvasive PD cannot be considered superior or standardat this time. Randomized controlled trials or prospectivecohort studies, which avoid selection and experimenterbias and control for confounding factors are necessaryto adequately evaluate this question before routine appli-cation can be recommended.

Author Contributions

Study conception and design: Correa-Gallego, Dinkelspiel,Sulimanoff, Fisher, Vinuela

Acquisitionof data:Correa-Gallego,Dinkelspiel, Sulimanoff,Vinuela

Analysis and interpretation of data: Correa-Gallego,Dinkelspiel, Fisher, Vinuela, Kingham, Fong,DeMatteo, D’Angelica, Jarnagin, Allen

Drafting of manuscript: Correa-Gallego, Dinkelspiel,Sulimanoff, Fisher, Vinuela

Critical revision: Correa-Gallego, Dinkelspiel, Sulimanoff,Fisher, Vinuela, Kingham, Fong, DeMatteo,D’Angelica, Jarnagin, Allen

REFERENCES

1. Clinical Outcomes of Surgical Therapy Study Group.A comparison of laparoscopically assisted and open colectomyfor colon cancer. N Engl J Med 2004;350:2050e2059.

2. Colon Cancer Laparoscopic or Open Resection Study Group,Buunen M, Veldkamp R, et al. Survival after laparoscopicsurgery versus open surgery for colon cancer: long-termoutcome of a randomised clinical trial. Lancet Oncol 2009;10:44e52.

3. Fleshman J, Sargent DJ, Green E, et al. Laparoscopic colec-tomy for cancer is not inferior to open surgery based on5-year data from the COST Study Group trial. Ann Surg2007;246:655e662; discussion 662e664.

4. Huscher CG, Mingoli A, Sgarzini G, et al. Laparoscopic versusopen subtotal gastrectomy for distal gastric cancer: five-year

results of a randomized prospective trial. Ann Surg 2005;241:232e237.

5. Kim HH, Hyung WJ, Cho GS, et al. Morbidity and mortalityof laparoscopic gastrectomy versus open gastrectomy for gastriccancer: an interim reportea phase III multicenter, prospective,randomized Trial (KLASS Trial). Ann Surg 2010;251:417e420.

6. Kitano S, Shiraishi N, Fujii K, et al. A randomized controlledtrial comparing open vs laparoscopy-assisted distal gastrectomyfor the treatment of early gastric cancer: an interim report.Surgery 2002;131:S306eS311.

7. Tozzi R, Malur S, Koehler C, et al. Laparoscopy versus lapa-rotomy in endometrial cancer: first analysis of survival ofa randomized prospective study. J Minim Invasive Gynecol2005;12:130e136.

8. Walker JL, Piedmonte MR, Spirtos NM, et al. Recurrence andsurvival after random assignment to laparoscopy versus lapa-rotomy for comprehensive surgical staging of uterine cancer:Gynecologic Oncology Group LAP2 Study. J Clin Oncol2012;30:695e700.

9. Biere SS, van Berge Henegouwen MI, Maas KW, et al. Mini-mally invasive versus open oesophagectomy for patients withoesophageal cancer: a multicentre, open-label, randomisedcontrolled trial. Lancet 2012;379:1887e1892.

10. Nix J, Smith A, Kurpad R, et al. Prospective randomizedcontrolled trial of robotic versus open radical cystectomy forbladder cancer: perioperative and pathologic results. EurUrol 2010;57:196e201.

11. Gagner M, Pomp A. Laparoscopic pylorus-preserving pancrea-toduodenectomy. Surg Endosc 1994;8:408e410.

12. Mabrut JY, Fernandez-Cruz L, Azagra JS, et al. Laparoscopicpancreatic resection: results of a multicenter European studyof 127 patients. Surgery 2005;137:597e605.

13. Ito M, Horiguchi A, Ishihara S, et al. Laparoscopic pancreaticsurgery: totally laparoscopic pancreatoduodenectomy andreconstruction. Pancreas 2009;38:1009.

14. Palanivelu C, Rajan PS, Rangarajan M, et al. Evolution intechniques of laparoscopic pancreaticoduodenectomy: a decadelong experience from a tertiary center. J Hepatobiliary Pan-creat Surg 2009;16:731e740.

15. Del Chiaro M, Moretto C, Croce C, et al. 34 robotic pancre-atectomies: short term results. Pancreatology 2010;10:289.

16. Kendrick ML, Cusati D. Total laparoscopic pancreaticoduode-nectomy: feasibility and outcome in an early experience. ArchSurg 2010;145:19e23.

17. Mackenzie S, Kosari K, Sielaff T. The robotic Whipple: initialexperience and technical considerations. HPB 2010;12:96e97.

18. Ammori BJ, Ayiomamitis GD. Laparoscopic pancreaticoduo-denectomy and distal pancreatectomy: a UK experience anda systematic review of the literature. Surg Endosc 2011;25:2084e2099.

19. Damrah O, Rahman SH. Early experience of total laparo-scopic pancreaticoduodenectomy. Pancreatology 2011;11:308.

20. Fischer CP, Fahy BN, Dunkin BJ, et al. Laparoscopic pancrea-ticoduodenectomy for cancer: margin status, adequacy of resec-tion and 90 day outcomes. Gastroenterology 2011;140:S991.

21. Stauffer JA, Goldberg RF, Parker JM, et al. Pancreatic fistularate is acceptable after laparoscopic pancreatic surgery. SurgEndosc Other Intervent Tech 2011;25:S332.

22. Zureikat AH, Bartlett DL, Steel JL, et al. Robotic-assistedpancreaticoduodenctomy at a high-volume pancreas center.J Clin Oncol 2011;29[suppl]:258.


23. Belluomini MA, De Lio N, Signori S, et al. Eighty laparoscopicrobot-assisted pancreatectomies. HPB 2012;14:259e260.

24. Belluomini MA, De Lio N, Signori S, et al. Total robotic pan-creaticoduodenectomy [abstract]. HPB 2012;14:245.

25. GumbsA,ZuckerN,GayetB.Laparoscopicpancreatic surgerywithrobotic assistance: 197 procedures performed with a robotically-controlled laparoscope holder [abstract]. HPB 2012;14:86.

26. Kamyab A, Khatlani K, Jacobs MJ. Total laparoscopic pan-creaticoduodenectomy: early single institution experience[abstract]. HPB 2012;14:38.

27. Zeh HJ, Zureikat AH, Secrest A, et al. Outcomes after robot-assisted pancreaticoduodenectomy for periampullary lesions.Ann Surg Oncol 2012;19:864e870.

28. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA state-ment for reporting systematic reviews and meta-analyses ofstudies that evaluate health care interventions: explanationand elaboration. Ann Intern Med 2009;151:W65eW94.

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

30. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring incon-sistency in meta-analyses. BMJ 2003;327:557e560.

31. Oxman AD, Guyatt GH. Validation of an index of the qualityof review articles. J Clin Epidemiol 1991;44:1271e1278.

32. Asbun HJ, Stauffer JA. Laparoscopic vs open pancreaticoduodenec-tomy:overall outcomes and severityof complicationsusing the accor-dion severity grading system. J Am Coll Surg 2012;215:810e819.

33. Buchs NC, Addeo P, Bianco FM, et al. Robotic versus openpancreaticoduodenectomy: a comparative study at a singleinstitution. World J Surg 2011;35:2739e2746.

34. Chalikonda S, Aguilar-Saavedra JR, Walsh RM. Laparoscopicrobotic-assisted pancreaticoduodenectomy: a case-matched compar-ison with open resection. Surg Endosc 2012;26:2397e2402.

35. Lai EC, YangGP, Tang CN. Robot-assisted laparoscopic pancrea-ticoduodenectomy versus open pancreaticoduodenectomy - Acomparative study. Int J Surg 2012;10:475e479.

36. Zhou NX, Chen JZ, Liu Q, et al. Outcomes of pancreatoduo-denectomy with robotic surgery versus open surgery. Int J MedRobot 2011;7:131e137.

37. Zureikat AH, Breaux JA, Steel JL, et al. Can laparoscopic pan-creaticoduodenectomy be safely implemented? J GastrointestSurg 2011;15:1151e1157.

38. Higgins JP, Altman DG, Gotzsche PC, et al. The CochraneCollaboration’s tool for assessing risk of bias in randomisedtrials. BMJ 2011;343:d5928.

39. Reddy JR, Saxena R, Singh RK, et al. Reoperation followingpancreaticoduodenectomy. Int J Surg Oncol 2012;2012:218248.

40. Aimaq R, Akopian G, Kaufman HS. Surgical site infectionrates in laparoscopic versus open colorectal surgery. Am Surg2011;77:1290e1294.

41. Varela JE, Wilson SE, Nguyen NT. Laparoscopic surgerysignificantly reduces surgical-site infections compared withopen surgery. Surg Endosc 2010;24:270e276.

42. Balcom JH, Rattner DW, Warshaw AL, et al. Ten-year expe-rience with 733 pancreatic resections: changing indications,older patients, and decreasing length of hospitalization. ArchSurg 2001;136:391e398.

43. Fernandez-del Castillo C, Rattner DW, Warshaw AL. Stan-dards for pancreatic resection in the 1990s. Arch Surg 1995;130:295e299; discussion 299e300.

44. Winter JM, Cameron JL, Campbell KA, et al. 1423 pancrea-ticoduodenectomies for pancreatic cancer: a single-institutionexperience. J Gastrointest Surg 2006;10:1199e1210; discus-sion 1210e1211.

45. Winter JM, Brennan MF, Tang LH, et al. Survival after resec-tion of pancreatic adenocarcinoma: results from a single insti-tution over three decades. Ann Surg Oncol 2012;19:169e175.

46. Adsay NV, Basturk O, Altinel D, et al. The number of lymphnodes identified in a simple pancreatoduodenectomy spec-imen: comparison of conventional vs orange-peeling approachin pathologic assessment. Mod Pathol 2009;22:107e112.

47. Tsui C, Klein R, Garabrant M. Minimally invasive surgery:national trends in adoption and future directions for hospitalstrategy. Surg Endosc 2013;27:2253e2257.

48. Barbash GI, Glied SA. New technology and health care costsethe case of robot-assisted surgery. N Engl J Med 2010;363:701e704.


minimally-invasive vs open pancreaticoduodenectomy...

Documents