Journal of Visceral Surgery (2011) 148, e279—e286

. B

tivetive,tive,tive,

011

PRESS REVIEW

Press review issue 04

C. Mariettea,∗, S

a Service de chirurgie digesb Service de chirurgie digesc Service de chirurgie digesd Service de chirurgie diges

Available online 12 August 2

Effect of early oral feeding after gastric cancersurgery: a result of randomized clinical trial

� Hur H, Kim SG, Shim JH, et al. Surgery

2011;149:561—8. DOI:10.1016/j.surg.2010.10.003

BackgroundTo date, early oral feeding after gastrectomy for gastric

cancer has not been accepted universally. Therefore, weperformed a randomized clinical trial to determine whetherearly oral feeding after curative surgery for gastric cancercan be tolerated and whether it has an effect on recovery.Methods

From July 2008 to February 2009, 58 patients wereenrolled and four were excluded according to set criteria.The patients in the early feeding group began a liquid diet onthe second postoperative day, and then were fed a soft dietfrom the third day until the day they were discharged. Thepatients in the control group began a liquid diet on the fourthday. The primary endpoint of this study was the duration ofpostoperative hospitalization.Results

No significant differences were found in the clinico-operative characteristics between the two groups. Theduration of hospitalization (P = 0.044) and time until flatus(P = 0.036) in the early group were decreased significantly.With regard to the rates of morbidity, cost of hospitalization,postoperative symptoms, and pain scales, no significant

∗ Corresponding author.E-mail address: c-mariette@chru-lille.fr (C. Mariette).

1878-7886/$ — see front matterdoi:10.1016/j.jviscsurg.2011.06.004

enoistb, D. Goerec, M. Pocardd

et générale, hôpital Huriez, CHRU de Lille, Lille, FranceCHU Le Kremlin-Bicêtre, Le Kremlin-Bicêtre, Franceinstitut Gustave-Roussy, Villejuif, Francehôpital Lariboisière, Paris, France

differences were found. The quality of life scores weredecreased significantly at the fatigue (P = 0.007) and nau-sea and vomiting (P = 0.048) immediately after operation inthe early feeding group.Conclusion

Early oral feeding after gastric cancer surgery is feasible

and can result in shorter hospitalization and improvementsin several aspects of quality of life in the early postoperativeperiod.Comments1. Based on the belief that increased intraluminal pressure

might cause anastomotic leak, early oral feeding is usu-ally not permitted after gastrectomy. This randomizedtrial shows that early oral feeding after gastric cancersurgery is feasible without increasing postoperative mor-bidity, in particular, leakage or aspiration; the benefitsinclude shorter hospitalization, improved quality of life,and quicker return to intestinal transit.

2. In accordance with several other studies [1,2], it alsoshows that a nasogastric tube is not necessary.

3. However, of note, this study does allow to say whetherearly oral feeding is feasible after total gastrectomy: thenumber of patients was small and the risk of leakage afteresophagojejunostomy is higher.

4. Again, because of the small number of patients, it wasnot possible to demonstrate the theoretical advantagesof early oral feeding on decreased morbidity, specificallyinfection [3] and improved healing [4].

5. The corollary to early oral feeding is the decreased needfor parenteral nutrition, reducing the related costs. Inmost patients undergoing partial gastrectomy who toler-

e

R[[[[

B

oAfHlM

lpeawtPeafR

aDrmbAcC

fprC1

2

3

4

5

R[[[[[[

O

oeD

S

280

ate early oral feeding, parenteral alimentation, with itsattendant risks (sepsis, thrombosis. . .) is not necessary.

eferences1] Eur J Surg 1994l;160:369—73.2] J Chir (Paris) 2005;142:14—28.3] J Parenteral Enteral Nutr 1997;21:202—7.4] Ann Surg 1974;180:765—72.

Delayed colo-anal anastomosis is an alternative toprophylactic diverting stoma after total mesorectalexcision for middle and low rectal carcinomas

� Jarry J, Faucheron JL, Moreno W, Bellera CA,Evrard S. Eur J Surg Oncol 2011;37:127—33.DOI:10.1016/j.ejso.2010.12.008

ackgroundAfter total mesorectal excision (TME), a low col-

rectal or colo-anal anastomosis is usually performed.prophylactic covering stoma is often used, especially

or patients receiving neoadjuvant chemo-radiotherapy.owever, morbidity is high, mainly due to anastomotic

eakage.ethods

From May 2000 to October 2008, patients with middle orow rectal cancer who underwent a trans-anal pull-throughrocedure after TME were prospectively recorded. No cov-ring stoma was performed in these patients. However, theyll underwent a delayed colo-anal anastomosis (DCA), whichas performed 6 days following the TME, on average. Both

he surgical technique and follow-up were standardized.atients with T3, T4 and/or N+ cancers were given preop-rative radiotherapy. A retrospective analysis was done tossess postoperative mortality, morbidity, and oncologic andunctional results.esults

One hundred consecutive patients with rectal tumors atmedian distance of 5 cm from the anal verge underwent

CA after TME. The 5-year overall and disease-free survivalates were 81% and 66%, respectively. The postoperativeortality rate was 3% and the overall post-operative mor-idity rate was 36%, with only three anastomotic leakages.

fter 2 years, 73% of the patients had good functional out-omes.onclusion

The trans-anal pull-through procedure after TME,ollowed by DCA seems to be a safe and efficient sphincter-reserving procedure to treat patients with middle or lowectal cancer while avoiding a prophylactic, diverting stoma.omments. Prophylactic diverting ileostomy is not a risk-free proce-

dure: related complications occur in nearly 10% of cases[5], quality of life is impaired [6] and increased costsincurred.

. The theoretical advantages of delayed colo-anal anasto-mosis (DCA) are:a. the resection can be performed via laparotomy or

laparoscopy, i.e. no incision is required to extract thespecimen in laparoscopy,

b. the duration of operation is shorter,c. the need to fashion an ileostomy is circumvented,

thereby avoiding related complications,d. the technique promotes early postoperative recov-

ery with less pain, less risk of leakage, and avoidsa separate hospital stay,

e. there is an overall, reduction of costs.

tP

icAm2fmI

gM

igmR

ftthe

C. Mariette et al.

. Potential disadvantages include suboptimal functionalresults related to the absence of a reservoir and sphinc-teric disorders.

. In this non-comparative retrospective study, most post-operative criteria were in favor of DCA. Functionalresults, on the other hand, are more difficult to evalu-ate because there was no control group; they seem to beinferior to those obtained when a reservoir is fashioned,especially during the first few postoperative years. Con-versely, postoperative complications are more frequentafter reservoir confection, which might also alter thefunctional outcome negatively.

. The authors currently recommend DCA routinely becauseof their favorable outcomes. Multicenter comparativestudies with larger number of patients are necessarybefore this technique can be generalized. DCA seems par-ticularly interesting in complex cases of pelvic surgery,and specially, in the case of chronic protracted pelvicfistula or morbidly obese patients.

eferences1] J Am Coll Surg 1997;185:105—13.2] Ann Surg 1988;207:61—5.3] Br J Surg 2001;88:400—4.4] Br J Surg 2003;90:1261—6.5] Br J Surg 2001;88:360—3.6] Dis Colon Rectum 1996;39:978—80.

Gastric bypass vs sleeve gastrectomy for type 2diabetes mellitus: A randomized controlled trial

� Lee WJ, Chong K, Ser KH, et al. Arch Surg2011;146:143—8. DOI:10.1001/archsurg.2010.326

bjectivesTo determine the efficacies of two weight-reducing

perations on diabetic control and the role of duodenumxclusion.esign

Double-blind randomized controlled trial.etting

Department of Surgery of the Min-Sheng General Hospi-al, National Taiwan University.

atients

We studied 60 moderately obese patients (body massndex > 25 and < 35) aged > 30 to < 60 years who had poorlyontrolled type 2 diabetes mellitus (T2DM) (hemoglobin[1c] [HbA(1c)] > 7.5%) after conventional treatment (> 6onths) from September 1, 2007, through June 30,

008. Patients and observers were masked during theollow-up, which ended in 2009, 1 year after final enroll-ent.

nterventionsGastric bypass with duodenum exclusion (n = 30) vs sleeve

astrectomy without duodenum exclusion (n = 30).ain outcome measures

The primary outcome was remission of T2DM (fast-ng glucose < 126 mg/dL and HbA[1c] < 6.5% withoutlycemic therapy). Secondary measures included weight andetabolic syndrome. Analysis was by intention to treat.esults

Of the 60 patients enrolled, all completed the 12-monthollow-up. Remission of T2DM was achieved by 28 (93%) inhe gastric bypass group and 14 (47%) in the sleeve gastrec-omy group (P = 0.02). Participants assigned to gastric bypassad lost more weight, achieved a lower waist circumfer-nce, and had lower glucose, HbA(1c), and blood lipid levels

Press review issue 04

than the sleeve gastrectomy group. No serious complicationsoccurred in either group.Conclusions

Participants randomized to gastric bypass were morelikely to achieve remission of T2DM. Duodenum exclusionplays a role in T2DM treatment and should be assessed.Comments1. There is consensual agreement that bariatric surgery is

indicated when the BMI > 35 kg/m2. Metabolic surgery,however, has been proposed recently for patients withlower BMI’s. The question here is to know whetherremission of Type 2 diabetes is secondary to simpleweight loss from restrictive surgery or to modificationsof hormonal secretion related to duodenal and proximaljejunal bypass.

2. This is the first randomized trial evaluating the valueof metabolic surgery in diabetics who are not mor-bidly obese: both procedures improve diabetes, with,however, a better control and better resolution ofthe metabolic syndrome in patients undergoing gastricbypass.

3. These results must be counter-balanced by the technicaldifficulty and increased morbidity associated with gastricbypass, compared to sleeve gastrectomy [1], especiallysince there were no data on hormonal secretion or long-term results in this study.

Reference[1] JAMA 2004;292:1724—1737.

Colonic stenting versus emergency surgery for acuteleft-sided malignant colonic obstruction: amulticentre randomized trial

� van Hooft JE, Bemelman WA, OldenburgB, et al. Lancet Oncol 2011;12:344—52.DOI:10.1016/S1470-2045(11)70035-3

BackgroundColonic stenting as a bridge to elective surgery is an

alternative for emergency surgery in patients with acutemalignant colonic obstruction, but its benefits are uncer-tain. We aimed to establish whether colonic stenting has

better health outcomes than does emergency surgery.Methods

Patients with acute obstructive left-sided colorectalcancer were enrolled from 25 hospitals in the Nether-lands and randomly assigned (1:1 ratio) to receive colonicstenting as a bridge to elective surgery or emergencysurgery. The randomization sequence was computer gen-erated with permuted blocks and was stratified by centre;treatment allocation was concealed by use of a web-basedapplication. Investigators and patients were unmasked totreatment assignment. The primary outcome was meanglobal health status during a 6-month follow-up, whichwas assessed with the QL2 subscale of the EuropeanOrganisation for Research and Treatment of Cancer quality-of-life questionnaire (EORTC QLQ-C30). Analysis was byintention to treat. This study is registered, numberISRCTN46462267.Findings

Between March 9, 2007, and August 27, 2009, 98 patientswere assigned to receive colonic stenting (n = 47 patients) oremergency surgery (n = 51). Two successive interim analysesshowed increased 30-day morbidity in the colonic stent-ing group, with an absolute risk increase of 0.19 (95% CI—0.06 to 0.41) in analysis of the first 60 patients (14 of

e281

28 patients receiving colonic stenting vs 10 of 32 receivingemergency surgery), and an absolute risk increase of 0.19(—0.01 to 0.37) in analysis of the first 90 patients (23 of 47patients vs 13 of 43). In accordance with the advice of thedata safety monitoring committee, the study was suspendedon September 18, 2009, and ended on March 12, 2010. Atthe final analysis of 98 patients, mean global health statusduring follow-up was 63.0 (SD 23.8) in the colonic stentinggroup and 61.4 (SD 21.9) in the emergency surgery group;after adjustment for baseline values, mean global healthstatus did not differ between treatment groups (—4.7, 95% CI—14.8 to 5.5, P = 0.36). No difference was recorded betweentreatment groups in 30-day mortality (absolute risk dif-ference —0.01, 95% CI —0.14 to 0.12, P = 0.89), overallmortality (—0.02, —0.17 to 0.14, P = 0.84), morbidity (—0.08,—0.27 to 0.11, P = 0.43), and stoma rates at latest follow-up (0.09, —0.10 to 0.27, P = 0.35). However, the emergencysurgery group had an increased stoma rate directly afterinitial intervention (0.23, 0.04 to 0.40, P = 0.016) and areduced frequency of stoma-related problems (between-group difference —12.0, —23.7 to —0.2, P = 0.046). The mostcommon serious adverse events were abscess (three in thecolonic stenting group vs four in the emergency surgerygroup), perforations (six vs none), and anastomotic leak-age (five vs one), and the most common adverse eventswere pneumonia (three vs one) and wound infection (onevs three).Interpretation

Colonic stenting has no decisive clinical advantages toemergency surgery. It could be used as an alternative treat-ment in as yet undefined subsets of patients, although withcaution because of concerns about tumor spread caused byperforations.Comments1. This trial was discontinued at the time of intermediary

analysis because of the high rate of complications in thestent arm, confirming the detrimental effect recentlyreported in a French study, which had also shown thatstent insertion was not able to decrease the rate ofstomas [1].

2. In the stent group, complications occurred mainly dur-

ing operation following stent insertion; there was a 25%anastomotic fistula rate, underscoring the difficulties inresection incurred by previous stent placement. More-over, the unusually high rate of anastomotic fistula couldalso be related to the fact that all the patients under-went operation within 15 days of stent insertion; delayingsurgery a bit longer would undoubtedly have been prefer-able.

3. Pathology reported tumoral perforation in 10% ofpatients undergoing colonic resection after stent inser-tion. The long-term carcinologic consequences ofperforation are not known, and were not evaluated inthis study. Future studies must absolutely address thisparticular point.

References[1] Surg Endosc 2011;25:14—21.[2] Arch Surg 2009;144:1127—32.

Increased local recurrence and reduced survival fromcolorectal cancer following anastomotic leak:systematic review and meta-analysis

� Mirnezami A, Mirnezami R, Chandraku-maran K, et al. Ann Surg 2011;253:890—9.DOI:10.1097/SLA.0b013e3182128929

e

O

mcusB

ohriivM

tbisDrwaR

drtrwioCrLaC

aacbC1

2

3

R[[

B

ittecM

tinApRecrtttstsR

w(cC

uoa

282

bjectiveTo examine the long-term oncological impact of anasto-

otic leakage (AL) after restorative surgery for colorectalancer using meta-analytical methods. Outcomes eval-ated were local recurrence, distant recurrence, andurvival.ackground

Recurrence after potentially curative surgery for col-rectal cancer remains a significant clinical problem andas a poor prognosis. AL may be a risk factor for diseaseecurrence, however available studies have been conflict-ng. A meta-analysis was conducted to investigate thempact of AL on disease recurrence and long-term sur-ival.ethods

Studies published between 1965 and 2009 evaluatinghe long-term oncological impact of AL were identifiedy an electronic literature search. Outcomes evaluatedncluded local recurrence, distant recurrence, and cancerpecific survival. Meta-analysis was performed using theerSimonian-Laird random-effects model to compute oddsatio and 95% confidence intervals. Study heterogeneityas evaluated using Q statistics and I and publication biasssessed with funnel plots and Egger’s test.esults

Twenty-one studies comprising 13 prospective nonran-omized studies, one prospective randomized, and sevenetrospective studies met the inclusion criteria, yielding aotal of 21,902 patients. For rectal anastomoses, the oddatios (OR) of developing a local recurrence when thereas AL was 2.05 (95% CI = 1.51—2.8; P = 0.0001). For stud-

es describing both colon and rectal anastomoses, the ORf local recurrence when there was an AL was 2.9 (95%I = 1.78—4.71; P < 0.001). The OR of developing a distantecurrence after AL was 1.38 (95% CI = 0.96—1.99; P = 0.083).ong-term cancer specific mortality was significantly higherfter AL with an OR of 1.75 (95% CI = 1.47—2.1; P = 0.0001).onclusions

AL has a negative prognostic impact on local recurrencefter restorative resection of rectal cancer. A significantssociation between colorectal AL and reduced long-term

ancer specific survival was also noted. No associationetween AL and distant recurrence was found.omments. This is the first meta-analysis on this topic, confirm-

ing that anastomotic leak (AL) impacts long-term resultsafter colorectal surgery for cancer. Methodologicallyspeaking the robustness of this meta-analysis is weaksince only one controlled trial was included. Moreover,the studies were heterogeneous in quality, especially asconcerns the definition of AL.

. The clinical impact of this study is finally somewhatweak. Since AL is associated with worse prognosis, itmight be interesting to propose adjuvant treatmentto these patients (chemo or radiation therapy). Theproblem is that the general condition of patients withpostoperative AL is usually precarious, and does not allowadministration of such treatment.

. Separate analysis of symptomatic and asymptomatic(radiologic only) LA would have been of interest becauseit is probable that asymptomatic LA without pelvic sepsisimpacts less on local recurrence.

eferences1] Br J Surg 2003;90:1261—6.2] Br J Surg 2009;96:1066—75.

C1

2

3

C. Mariette et al.

Ex vivo technical skills training transfers to theoperating room and enhances cognitive learning: arandomized controlled trial

� Palter VN, Grantcharov T, Harvey A, et al.Ann Surg 2011;253:886—9. DOI:10.1097/SLA.0b013e31821263ec

ackgroundSurgical training in the operating room includes acquir-

ng technical skills and cognitive knowledge. Technical skillsraining on simulated models has been shown to improveechnical performance in the operating room, and may alsonhance the acquisition of other skills by freeing cognitiveapacity. This has yet to be investigated.ethods

We conducted a single-blinded randomized controlledrial to assess the effect of ex vivo technical skills train-ng on cognitive learning in the operating room. Eighteenovice surgical residents were randomized to two groups.ll participants were taught the basics of fascial closure anderformed one closure on a low fidelity synthetic model.esidents in the intervention group practiced on the mod-ls until technical proficiency was reached. Residents in theontrol group had no further contact with the models. Allesidents then performed a fascial closure on a patient inhe operating room while listening to a script that con-ained relevant clinical information. A validated evaluationool was used to assess the technical merit of the clo-ure. Finally, all participants completed a multiple-choiceest designed to test the information retained from thecript.esults

The technical performance of the ex vivo trained groupas significantly higher than that of the untrained group

P = 0.04). The ex vivo trained group also performed signifi-antly better on the cognitive retention test (P = 0.03).onclusions

Technical skills training using a low fidelity synthetic sim-lator resulted in improved technical performance in theperating room, and enhanced the ability of residents tottend to cognitive components of surgical expertise.

omments. This original study is the first to evaluate the impact

of simulation on theoretical knowledge acquisition insurgery.

. It is surprising that the technical performance in thetrained group of residents did not improve: the medianOSATS score was the same before and after the ex-vivo training sessions. This leads us to think that,notwithstanding stress related to the examination, theperformance of fascial closure in the human is more com-plex than in the simulator exercise.

. Several biases intervene in the interpretation of results.First, patients were different for each day of examinationfor each resident, and it is therefore possible that certainpatients were more difficult than others to close. More-over, in the trained group, the period between the endof training and the beginning of examination was vari-able and the effects of training could have waned withtime in certain candidates. Last, how can we be surethat some theoretical notions concerning fascial closurewere not communicated orally to the residents, duringthe exercise, facilitating a higher score in the theoret-ical post-fascial closure test. It might have been morejudicious to read a script that had nothing to do withfascial closure.

Press review issue 04

References[1] Ann Surg 2002;236:458—63.[2] Ann Surg 2005;241:364—72.

External pancreatic duct stent decreases pancreaticfistula rate after pancreaticoduodenectomy:prospective multicenter randomized trial

� Pessaux P, Sauvanet A, Mariette C,et al. Ann Surg 2011;253:879—88. DOI10.1097/SLA.0b013e31821219af

ObjectivePancreatic fistula (PF) is a leading cause of morbidity and

mortality after pancreaticoduodenectomy (PD). The aim ofthis multicenter prospective randomized trial was to com-pare the results of PD with an external drainage stent versusno stent.Methods

Between 2006 and 2009, 158 patients who underwentPD were randomized intraoperatively to either receive anexternal stent inserted across the anastomosis to drainthe pancreatic duct (n = 77) or no stent (n = 81). The cri-teria of inclusion were soft pancreas and a diameter ofWirsung < 3 mm. The primary study end point was PF ratedefined as amylase-rich fluid (amylase concentration greaterthan three times the upper limit of normal serum amy-lase level) collected from the peripancreatic drains afterpostoperative day 3. CT scan was routinely done on day7.Results

The two groups were comparable concerning demo-graphic data, underlying pathologies, presenting symptoms,presence of comorbid illness, and proportion of patientswith preoperative biliary drainage. Mortality, morbidity,and PF rates were 3.8%, 51.8%, and 34.2%, respectively.Stented group had a significantly lower overall PF (26% vs42%; P = 0.034), morbidity (41.5% vs 61.7%; P = 0.01), anddelayed gastric emptying (7.8% vs 27.2%; P = 0.001) ratescompared with nonstented group. Radiologic or surgicalintervention for PF was required in nine patients in thestented group and 12 patients in the nonstented group.

There were no significant differences in mortality rate (3.7%vs 3.9%; P = 0.37) and in hospital stay (22 days vs 26 days;P = 0.11).Conclusion

External drainage of pancreatic duct with a stentreduced. PF and overall morbidity rates after PDin high risk patients (soft pancreatic texture anda nondilated pancreatic duct). This study is regis-tered at http://www.clinicaltrials.gov: Clinical trial ID#NCT01068886.Comments1. This is the second controlled trial showing a significant

reduction in the pancreatic fistula (PF) rate after pan-creatoduodenectomy (PD) in patients with external stentdrainage of the main pancreatic duct. In the preced-ing study [2], patients were not selected according toparenchymal characteristics (soft or not), explaining thatthe PF rate was lower in the group without a stent (20%).However, the difference in PF between the two groupswas comparable, 13%. Because the efficacy of the stentwas confirmed in these two trials, external duct drainageshould be the standard for patients with soft parenchymaor when the Wirsung canal is < 3 mm in diameter.

e283

2. Other techniques have been tested to decrease the PFrate after PD:a. The type of pancreatodigestive anastomosis (gastric

or jejunal): three prospective trials [3—5] and onerecent meta-analysis [6],

b. Application of fibrin glue over the anastomosis [7],c. Main pancreatic duct closure without pancreaticoen-

terostomy [8]d. Technical variations in the pancreaticoenterostomy

[9,10]e. Administration of somatostatin analogue [11,12]. The

level of evidence in these reports is not high enoughto recommend any one technique to decrease the PFrate.

References[1] Surgery 2005;138:8—13.[2] Ann Surg 2007;246:425—433.[3] Ann Surg 1995;222:580—8.[4] Ann Surg 2005;242:767—71.[5] Am J Surg 2005;189:720—9.[6] Am J Surg 2007;193:171—83.[7] J Gastrointest Surg 2004;8:766—72.[8] Ann Surg 2003;237:57—65.[9] Am J Surg 1995;169:585—8.

[10] Br J Surg 2005; 92: 1059—67.[11] Ann Surg 2000;232:419—29.[12] HPB (Oxford) 2010;12:155—65.

Perioperative Chemotherapy Compared With SurgeryAlone for Resectable GastroesophagealAdenocarcinoma: An FNCLCC and FFCD MulticenterPhase III Trial

� Ychou M, Boige V, Pignon JP et al.J Clin Oncol 2011;29:1715—21.DOI:10.11200/JCO.2010.33.0597

PurposeAfter curative resection, the prognosis of gastroe-

sophageal adenocarcinoma is poor. This phase III trial wasdesigned to evaluate the benefit in overall survival (OS) ofperioperative fluorouracil plus cisplatin in resectable gas-

troesophageal adenocarcinoma.Patients and methods

Overall, 224 patients with resectable adenocarcinomaof the lower esophagus, gastroesophageal junction (GEJ),or stomach were randomly assigned to either perioper-ative chemotherapy and surgery (CS group; n = 113) orsurgery alone (S group; n = 111). Chemotherapy consistedof two or three preoperative cycles of intravenous cisplatin(100 mg/m[2]) on day 1, and a continuous intravenous infu-sion of fluorouracil (800 mg/m[2]/d) for five consecutivedays (day 1 to 5) every 28 days and three or four postop-erative cycles of the same regimen. The primary end pointwas OS.Results

Compared with the S group, the CS group had a better OS(5-year rate 38% vs 24%; hazard ratio [HR] for death: 0.69;95% CI, 0.50 to 0.95; P = 0.02); and a better disease-free sur-vival (5-year rate: 34% vs 19%; HR, 0.65; 95% CI, 0.48 to 0.89;P = 0.003). In the multivariable analysis, the favorable prog-nostic factors for survival were perioperative chemotherapy(P = 0.01) and stomach tumor localization (P < 0.01). Periop-erative chemotherapy significantly improved the curativeresection rate (84% vs 73%; P = 0.04). Grade 3 to 4 toxic-

e

ipC

eucC1

2

3

4

5

R[

B

atclM

msibDoR

shawde

C

pC1

2

3

R[[

O

tbB

ooeiM

ct

284

ty occurred in 38% of CS patients (mainly neutropenia) butostoperative morbidity was similar in the two groups.onclusion

In patients with resectable adenocarcinoma of the lowersophagus, GEJ, or stomach, perioperative chemotherapysing fluorouracil plus cisplatin significantly increased theurative resection rate, disease-free survival, and OS.omments. Ever since the results of the MAGIC trial were published

[1], patients with operable adenocarcinoma of the stom-ach receive perioperative chemotherapy (three pre andthree postoperative sessions) of combined epirubicin,cisplatin, and fluorouracil. The results of this trial, whileconfirming those of the MAGIC trial, make cisplatin-5FUa therapeutic option for patients with operable esopha-gogastric adenocarcinoma.

. However, the positive effect on survival seems relatedmore to preoperative (easier to administer) than post-operative (only 23% of patients received the entire dosepostoperatively) chemotherapy.

. One drawback of this trial is the slow accrual of patients(over 8 years), which led to discontinuation of the trialbefore reaching the total number of patients necessary.

. Even more original in this study, the results seem toindicate that preoperative chemotherapy is effectivein patients with lower esophageal and esophagogastricjunction adenocarcinoma, 75% of the patients recruitedin this study.

. Chemotherapy remains debated in patients with signetcell adenocarcinoma since the publication of the resultsof the retrospective multicenter study from Lille in theSFCD-ACHBT 2010 meeting. A multicenter randomizedstudy financed by the PHRC will soon begin.

eference1] N Engl J Med. 2006;355:11—20.

Surgical management of the duodenal manifestationsof familial adenomatous polyposis

� Parc Y, Mabrut JY, Shields C et al., on behalf

of the Mallorca Group. Br J Surg 2011;98:480—4.DOI:10.1002/bjs.7374

ackgroundDuodenal adenomas develop in patients with familial

denomatous polyposis, incurring a risk of carcinoma. Whenhis risk is high, surgery is indicated. The choice of surgi-al treatment can be difficult as evidence-based data areacking.ethods

This is a systematic review of the literature on the non-edical management of duodenal lesions arising in the

etting of familial adenomatous polyposis. Studies weredentified through searching MEDLINE. Studies publishedetween January 1965 and October 2009 were included.ata regarding number of subjects, complications, lengthf follow-up, recurrence rate and outcome were extracted.esults

Transduodenal resection does not differ from an endo-copic approach in terms of recurrence. Ampullectomyas limited application as only papillary lesions aremenable to treatment in this manner. Duodenectomyith pancreas preservation is preferable to pancreatico-uodenectomy unless malignancy is present, or cannot bexcluded.

ssrccswPcR

riddtsp((C

ict

C. Mariette et al.

onclusionSurgery should be reserved for advanced or malignant

olyps.omments. Patients with familial adenomatous polyposis (FAP) often

develop duodenal neoplasms underscoring the valueof routine endoscopic screening, privileging endoscopicresections (even repeated) at an early Spigelman stage[2].

. These results underline the value of management in spe-cialized centers, where surgeons and gastroenterologistscan propose optimal therapeutic strategies.

. It is however difficult to establish evidence-based rec-ommendations because the level of evidence in thesereports, although originating most often from expert cen-ters, is low (level 4).

eferences1] J Chir (Paris) 2002;139:260—7.2] Lancet 1989;2:783—5.

The Conservative management of severe causticgastric injuries

� Zerbib P, Voisin B, Truant S, et al. Ann Surg 2011;253:684—8. DOI:10.1097/SLA.0b013e31821110e8

bjectiveTo determine the safety of a conservative approach to

reating severe caustic injury in patients lacking clinical andiochemical signs of transmural necrosis.ackground

Esophagogastrectomy is thought to limit the progressionf severe caustic injury in the upper gastrointestinal tractbserved upon initial endoscopic examination. However,ndoscopic evaluation of the depth and spread of necrosiss challenging and may lead to unnecessary gastrectomy.ethods

From January 2002 to December 2008, 70 patients werelassified as having stage III gastric injury in an initial diges-ive tract endoscopic examination. When patients had no

igns of peritonitis, their treatment was determined byix clinical and biochemical factors of severity (abdominalebound tenderness, neuropsychiatric troubles, cardiovas-ular shock, metabolic acidosis, disseminated intravascularoagulation, and kidney failure) in addition to endoscopictaging. If one of these clinical and biochemical factorsas present, the patient underwent emergency laparotomy.atients with isolated stage III gastric injury were kept underlose observation.esults

Twenty-four of the 70 endoscopic stage III patientsequired emergency surgery. Conservative treatment wasnitiated in the remaining 46. There were four postoperativeeaths (5.7%). Fifteen patients required subsequent surgery:istal gastrectomy with Billroth I anastomosis (n = 7) for dis-al stricture and esophagoplasty for nondilatable esophagealtricture (n = 8). At the end of the follow-up period, total orartial gastric conservation was achieved in all 46 patients65.7%) and the esophagus was conserved in 38 patients54.3%).onclusion

In the absence of clinical and biological signs of sever-ty, conservative management of stage III gastric injury islinically feasible, precludes gastrectomy and has a low mor-ality rate.

Press review issue 04

Comments1. This innovative article presents arguments in favor of

therapeutic de-escalation and organ preservation insevere caustic injury.

2. Although surgeons usually take into account theesophageal involvement, the authors here invite toreverse the rationale and consider above all the aspect ofthe gastric mucosa for decision-making. They show thatit is possible to preserve the stomach in 65% of cases.

3. However, the organ preservation policy leads to logicallypose the question of long-term function. These data arenot provided in this paper.

Efficacy of stapler versus hand-sewn after distalpancreatectomy (DISPACT): a randomized, controlledmulticentre trial

� Diener MK, Seiler CM, Rossion I, et al.Lancet 2011;377:1514—22. DOI:10.1016/S0140-6736(11)60237-7

BackgroundThe ideal closure technique of the pancreas after distal

pancreatectomy is unknown. We postulated that standard-ized closure with a stapler device would prevent pancreaticfistula more effectively than would a hand-sewn closure ofthe remnant.Methods

This multicentre, randomized, controlled, parallelgroup-sequential superiority trial was done in 21 Europeanhospitals. Patients with diseases of the pancreatic bodyand tail undergoing distal pancreatectomy were eligible andwere randomly assigned by central randomisation beforeoperation to either stapler or hand-sewn closure of the pan-creatic remnant. Surgical performance was assessed withintraoperative photo documentation. The primary endpointwas the combination of pancreatic fistula and death untilpostoperative day 7. Patients and outcome assessors weremasked to group assignment. Interim and final analysis wereby intention to treat in all patients in whom a left resectionwas done. This trial is registered, ISRCTN18452029.Findings

Between Nov 16, 2006, and July 3, 2009, 450 patientswere randomly assigned to treatment groups (221 stapler;229 hand-sewn closure), of whom 352 patients (177 stapler,175 hand-sewn closure) were analyzed. Pancreatic fistularate or mortality did not differ between stapler (56 [32%] of177) and hand-sewn closure (49 [28%] of 175; OR 0.84, 95% CI0.53—1.33; P = 0.56). One patient died within the first 7 daysafter surgery in the hand-sewn group; no deaths occurredin the stapler group. Serious adverse events did not differbetween groups.Interpretation

Stapler closure did not reduce the rate of pancreaticfistula compared with hand-sewn closure for distal pan-createctomy. New strategies, including innovative surgicaltechniques, need to be identified to reduce this adverseoutcome.Comments1. This study, with the highest level of evidence of studies,

evaluates the impact of these two techniques in distalpancreatectomy; no difference in outcome was found infavor of stapling.

2. Certain biases exist, however. The authors try, uncon-vincingly, to justify why randomization was not per-formed in the operation room. Of note, randomizing

e285

before the patient undergoes operation leads to guess-ing what the texture of the parenchyma will be, based onthe indication of the pancreatectomy and then to stratifythe patients accordingly. The actual (as opposed to thesupposed) consistency of the pancreatic parenchyma isnot clearly stated in the article.

3. The results of the meta-analysis presented in this article,including the results of the present trial, do not allowany conclusion as to which is the best manner to closethe pancreatic stump in distal pancreatectomy. Nonethe-less, stapling is the most widely used technique used inlaparoscopic surgery.

4. The value of identifying preventive measures is unde-niable. The randomized study by the Fédération deRecherche EN Chirurgie (FRENCH), evaluating the use ofthe Tachosil® sponge in this indication, may be able toprovide the answer.

Reference[1] Ann Surg 1999;229:693—98.

Outcome and cost analysis of sacral nerve modulationfor treating urinary and/or fecal incontinence

� Leroi AM, Lenne X, Dervaux B,et al. Ann Surg 2011;253:720—32.DOI:10.1097/SLA.0b013e318210f1f4

BackgroundSacral nerve modulation (SNM) is an established treat-

ment for urinary and fecal incontinence in patients for whomconservative management has failed.Objective

This study assessed the outcome and cost analysis of SNMcompared to alternative medical and surgical treatments.Methods

Clinical outcome and cost-effectiveness analyses wereperformed in parallel with a prospective, multicenter cohortstudy that included 369 consecutive patients with urgeurinary and/or fecal incontinence. The duration of follow-up was 24 months, and costs were estimated from thenational health perspective. Cost-effectiveness outcomeswere expressed as incremental costs per 50% of improved

severity scores (incremental cost-effectiveness ratio).Results

The SNM significantly improved the continence status(P < 0.005) and quality of life (P < 0.05) of patients withurge urinary and/or fecal incontinence compared to alter-native treatments. The average cost of SNM for urgeurinary incontinence was D8525 (95% confidence interval,D6686—D10,364; P = 0.001) more for the first 2 years com-pared to alternative treatments. The corresponding increasein cost for subjects with fecal incontinence was D6581 (95%confidence interval, D2077—D11,084; P = 0.006). When animprovement of more than 50% in the continence severityscore was used as the unit of effectiveness, the incrementalcost-effectiveness ratio for SNM was D94,204 and D185,160at 24 months of follow-up for urinary and fecal incontinence,respectively.Conclusions

The SNM is a cost-effective treatment for urge urinaryand/or fecal incontinence.Comments1. A systematic Cochrane review, published in 2007, con-

cluded that SNM was probably efficacious in selectedpatients with fecal incontience based on two studiesincluding a total of only 36 patients; this underscores the

e

2

286

lack of data addressing this topic [1]. The present studyis important because it is prospective, includes a largenumber of patients, and has a higher level of evidence toanswer the question of efficacy, with a follow up of twoyears.

. However, doubt persists as to the selection of the cohortindicating that the control group is not strictly compara-ble to the patients treated with SNM because of a largenumber of patients lost to follow-up. Moreover, ideally,the control group should not be the group of patientsin whom the treatment did not work. Notwithstandingthese methodological remarks, the study is important

3

R[

C. Mariette et al.

because a randomized study on the topic would beextremely difficult to perform.

. The message therefore is that SNM is efficacious forfecal incontinence, even when associated with urinaryincontinence and that it improves the quality of life ofpatients. Thanks to this study, SNM is now reimbursedin France as long as it is performed according to theFrench Health technology (HAS) criteria, i.e. includingat least a colorectal surgeon in the team performing theprocedure.

eference1] Cochrane Database Syst Rev 2010;(9):CD001757.