Laparoscopic Adjustable Silicone Gastric Banding Versus Vertical Banded Gastroplasty in Morbidly Obese Patients

LETTERS TO THE EDITOR

Prognostic FactorsFollowing Curative

Resection forPancreatic

Adenocarcinoma

To the Editor:The conclusion of this study is

drawn from a population-based analysisof 396 patients. The inclusion was basedupon the code for operation radical pan-creaticoduodenectomy or Whipple pro-cedure, which also included total pan-createctomy, proximal pancreatectomy,partial pancreatectomy and distal pan-createctomy. This means that it includesa wide range of pancreatic cancers. Inaddition, there is no information aboutthe specific kind of the tumor in case ofa process in the pancreatic head region,which in every series of Whipple resec-tion will include a substantial number ofampullary carcinomas and distal bileduct carcinomas. Furthermore, a centralpathologist should specifically look forneuroendocrine tumors. Lacking this in-formation and considering the variationin the types of tumor, it is impossible tocome to a valid conclusion. This is sup-ported by the fact that the survival rate isextremely high. The 5-year survival rateis depicted in the figures above 30%,median survival time is nearly 18months, which has to mean that tumorslike distal bile duct cancer and ampul-lary cancer are included in the series andmay be also endocrine tumors.

As I have said above, this alsomeans that conclusions about the prog-nostic factors cannot be made. The titleof the article is misleading, for it sug-gests that prognostic factors be given forpancreatic adenocarcinoma, which usu-ally means carcinoma of the head of thepancreas instead of a variation of can-cers. We know that the prognosis ofcancer in corpus or tail of the pancreas isworse. On the other hand the prognosisof cyst adenocarcinoma is better. Wealso know that the prognosis of distal

bile duct and peri-ampullary cancer ismuch better. Thus, conclusions aboutprognostic factors cannot be made fromthis study.

One of the most significant find-ings was the adjuvant combined chemo-radiotherapy as a prognostic factor. Thisis in contradiction with the evidence thatexists in the literature. We live in the eraof evidence-based medicine. Evidenceshould nowadays be based upon level Ior level II studies and there are only 2large prospective randomized studiesthat can be taken into account, which arethe studies of Klinkenbijl et al (AnnSurg 1999; 230:776-782) and Neoptolo-mos et al (Lancet, November 2001).These 2 large randomized studiesclearly indicate that there is no signifi-cant effect of chemo-radiotherapy inpancreatic cancer. The study in the Lan-cet is not even mentioned in the article!

To my surprise these 2 studies aremisinterpreted by the authors. Theywrite that: “two prospective randomizedclinical trials are in progress,” “althoughpreliminary data from the EORTCshowed no significant benefit of adju-vant-radiotherapy or chemotherapy on2-year survival, additional data from thisongoing as well as those from the ES-PAC-1 trial, should go a long way to-ward elucidating the role of adjuvantchemo-radiotherapy in the treatment ifpancreatic cancer patients.” This is aserious misinterpretation, these trials arenot ongoing trials and provide a firmstatement with good statistical evidence.Both trials are closed and finished.

The studies that are mentioned bythe authors that are consistent with theirfindings are all retrospective analysis, ex-cept for the historic GITSG-trial, whichhad a insufficient, very small number ofpatients.

Hans Jeekel, MD, PhDErasmus University

Rotterdam, The [email protected]

Are Actual StandardFluid Regimens in

Major Surgery Safe?

To the Editor:The paper by Brandstrup et al1 is

the first clinical trial to convincinglyshow that standard fluid replacementmethods in colorectal surgery are asso-ciated with increased morbidity andmortality. Long ago, fluid and salinesolutions were carefully administeredduring the antidiuretic perioperativephase of surgery, to maintain balanceand prevent weight gain.2,3 However,these cornerstone ideas were somehowreplaced in the last decades by the par-adigm that surgical patients need excep-tional high volume of fluids, irrespec-tively of measured requirements. Therecommended standard fluid replace-ment4 seems to drive from principles ofgoal-directed therapy5 aimed at increas-ing cardiac output in high-risk surgicalpatients. However, the consequences ofextrapolating these high requirements tomajor surgery were not, until now, as-sessed by clinical trials. Moreover, thereis a surprising refusal in surgical andintensive care trainees to admit that ex-cess perioperative fluid could be one ofthe underlying mechanisms of cardio-pulmonary postoperative complications.Paradoxically, despite the lack of thestudies supporting the benefit of the rec-ommended high fluid replacement, ac-tual researchers have to demonstrate thatthe old physiologic approach targetedtoward maintenance of body weight isassociated with a better outcome.

This paper also shows that thestandard group received a greater fluidvolume only during surgery and the firstpostoperative day, however, bodyweight was maintained higher in thisgroup for 6 days. It must be pointed outthat even the restricted group underwentan important positive fluid balance (notdescribed in the paper) revealed by theweight gain (also maintained 5 days)respect the weight recorded the morning

Annals of Surgery • Volume 240, Number 2, August 2004384

before surgery. Moreover, the authorsays nothing about the weight of neitherthe resected colon nor the expected 200-300g daily loss related to postoperativecatabolism. These issues probably con-ceal the real weight gain in both groupsand stress the inability of the kidneys toget rid of postoperative fluid excess. Thelatter findings are in close agreement withour own experience in medium complex-ity surgery6 and gives strong emphasizesto the growing risk of fluid overload bythe end of the first postoperative week.Brandstrup’s outstanding study might bethe starting point to answer the basic ques-tion: what are fluid requirements duringand after major surgery.

Miguel A. Jorge, MDHospital de Clinicas

Buenos Aires, [email protected]

REFERENCES1. Brandstrup B, Tønnesen H, Beier-Holgersen R,

et al. Effects of intravenous fluid restriction onpostoperative complications. Comparison oftwo perioperative fluid regimens: a randomizedassessor-blinded multicenter trial. Ann Surg.2003;238:641–646.

2. Moore FD. Metabolic care of the surgical pa-tient. Philadelphia: WB Saunders Co. 1959.

3. Tindall SF, Clark RG. The influence of highand low sodium intake on postoperative antidi-uresis. Br J Surg. 1981;68:639–644.

4. McKinlay S, Gan TJ. Intraoperative fluid man-agement and choice of fluids. In ASA Re-fresher Courses in Anesthesiology. EdSchwartz AJ. 2003;31:127–137.

5. Shoemaker WC, Appel PL, Kram HB, et al.Prospective trial of supranormal values of sur-vivors as therapeutic goals in high risk surgicalpatients. Chest. 1988;94:1176–1186.

6. Lopez Gaston O, Jorge MA, Basaluzzo JM, etal. Water and sodium requirements in the earlypostoperative period: comparison of two post-operative fluid regimens in medium complexitysurgery. Rev Arg Cir. 2001;80:119–124.

In Reply:On behalf of “The Danish Study

Group on Perioperative Fluid Therapy”I thank Dr. Jorge for the encouragingcomments on our work and for givingme this opportunity to discuss the liter-ature and report further details on thetrial.

Dr. Jorge points out the fact thatliberal fluid therapy in major electivesurgery has not been proven beneficialin clinical randomized trials and thatdata obtained from trials including pa-tients with traumatic injury, shock, sep-sis or other acute conditions should notbe extrapolated to patients undergoingelective surgery or vice versa. Dr. Jorgesuggests that standard fluid replace-ments may be influenced by the princi-ples of goal directed therapy, investigat-ing the effects of standard fluid therapy(not fluid restriction) versus standardfluid therapy plus additional fluid givento obtain a maximal cardiac output. Todiscuss these trials in a letter replywould not do them justice, but the veryidea of given fluid replacement to max-imal cardiac output, intriguing as it mayseem, is also to demand maximal workon the heart throughout surgery. Person-ally, I fail to see the potential benefit ofthis, especially in elderly patients. Themajority of the trials have tested theinfluence of fluid therapy in combina-tion with other therapy (ie, Dopexam-ine) and the results of these trials havenot been unanimous. Trials of goal-di-rected therapy investigating effects offluid therapy alone1–7 has ended up witha very small volume difference betweenthe groups on the day of operation(200–658 mL). With no registration ofthe fluid given on the surgical ward, theinterpretation of the results of these tri-als is most difficult. The most exhaus-tive and recent trial of goal directedtherapy in major surgery has failed toshow any superiority of the treatment,but on the contrary serious adverse ef-fects.8

In relation to our trial, Dr. Jorgepoints out the fact that in both groups aweight gain was maintained postopera-tively and suggest that this weight gainwould have been even greater if theweight of the removed colon and thekatabolic weight loss were considered. Iagree to this notion and to the observedinability of the kidneys to excrete post-operative fluid excess. This is important

because tissue edema consequently ismaintained for a long period of time. Ido not think, however, that the postop-erative weight gain in the restrictedgroup reflects a fluid overload. The pa-tients in both groups were allowed to eatand drink freely, but in addition theywere fed by tube to obtain a sufficientdaily caloric intake. Feeding was com-menced 4 hours postoperatively and 500mL of Nutriconcentrated® was given onthe day of surgery. In my opinion, thepostoperative weight increase observedin the restricted group was most likely aresult of feeding combined with postop-erative intestinal paralysis rather than agenuine fluid overload. Currently, how-ever, no simple and clinical feasiblemethod exists to differentiate betweenfluid in the intestines and fluid in thetissues, and I cannot prove my point ofview. Like Dr. Jorge, we find this datainteresting and a paper including furtherdata on fluid balance and body weight isin preparation.

Birgitte Brstrup MD, PhDGlostrup University Hospital

Glostrup, [email protected]

REFERENCES1. Conway DH, Mayall R, Abdul-Latif MS, et al.

Randomised controlled trial investigating theinfluence of intravenous fluid titration usingoesophageal Doppler monitoring during bowelsurgery. Anaesthesia. 2002;57:845–849.

2. Gan TJ, Soppitt A, Maroof M, et al.Goal-directed intraoperative fluid administra-tion reduces length of hospital stay after majorsurgery. Anesthesiology. 2002;97:820–826.

3. Mythen MG, Webb AR. Perioperative plasmavolume expansion reduces the incidence of gutmucosal hypoperfusion during cardiac surgery.Arch Surg. 1995;130:423–429.

4. Price J, Sear J, Venn R. Perioperative fluidvolume optimization following proximal fem-oral fracture (Cochrane review). The CochraneLibrary 2002.

5. Schultz RJ, Whitfield GF, Lamura JJ, et al. Therole of physiologic monitoring in patients withfractures of the hip. J Trauma. 1985;25:309–316.

6. Sinclair S, James S, Singer M. Intraoperativeintravascular volume optimisation and lengthof hospital stay after repair of proximal femoralfracture: randomised controlled trial. BMJ.1997;315:909–912.

Annals of Surgery • Volume 240, Number 2, August 2004 Letters to the Editor

© 2004 Lippincott Williams & Wilkins 385

7. Venn R, Steele A, Richardson P, et al.Randomized controlled trial to investigate in-fluence of the fluid challenge on duration ofhospital stay and perioperative morbidity inpatients with hip fractures. Br J Anaesth. 2002;88:65–71.

8. Sandham JD, Hull RD, Brant RF, et al. Arandomized, controlled trial of the use of pul-monary-artery catheters in high-risk surgicalpatients. New Engl J Med. 2003;348:5–14.

Effects of IntravenousFluid Restriction on

PostoperativeComplications:

Comparison of TwoPerioperative Fluid

Regimens:A Randomized

Assessor-BlindedMulticenter Trial

To the Editor:We read with great interest the

article by Dr. Birgitte Brandstrup andcolleagues1 in the November 2003 issueof the Annals of Surgery. Their objec-tive was to investigate the effects ofrestricted versus standard intravenousfluid regimens on complications aftercolorectal surgery. Having carefullystudied the paper we have several com-ments regarding the design and conclu-sions of the study.

Although this was a multicenterstudy they do not mention any protocolor guidelines for fluid management ap-plied during the postoperative period inthe standard regimen group. Further-more, they state that postoperative fluidadministration was given according tothe “standard” treatment of each hospi-tal and each ward. We find this a signif-icant flaw in the design of the study asmajor differences in fluid resuscitationcan occur between physicians or centers.When assessing the complication ratefollowing an intervention, it would beexpected to find some criteria for theprophylaxis measures and the treatment

of these complications. In the presentstudy, each 1 of the 8 hospitals used itsown routine for antibiotic and anti-thrombosis prophylaxis.

In addition, the researchers in-tended to include patients with ASAscore group I-III. However, due to theirexclusion criteria, 98% of patients in-cluded had an ASA score of I-II. Hence,most of their patients were relativelyhealthy. Our clinical experience taughtus that the major problems with fluidresuscitation occur in patients with ASAIII-IV.

Regarding the regimens: mainte-nance fluid support in the restrictedgroup during the surgical procedure andthe postoperative period consisted of 5%glucose. The recent trauma and criticalcare literature are not in favor of thiskind of therapy.2 In addition, the stan-dard group was treated with normal sa-line and HAES 6%, however, when therecommended dose of HAES wasreached albumin 5% was administered.The use of 5% albumin for fluid resus-citation is controversial, especially as itmay aggravate edema formation in areasof leaky capillaries.3,4 Hence, we haveto conclude that the regimen and typesof fluid used are not the currently prac-ticed nor advised.

The total amount of fluid admin-istered in the operative day ranged be-tween 1100 to 8050 mL in the restrictedgroup and between 2700 to 11,083 mLin the standard group. We noted thelarge range of fluids administered todifferent patients in each group. Basedon our experience, normally ASA I-IIpatients undergoing elective colecto-mies do not require such large amount offluids.

Although the randomization wascomputer generated, in the restrictedgroup most of the patients had an ileo-colic anastomosis contrary to a minorityin the standard group. We find it ofimportance to note that ileocolic anasto-moses are considered less prone to com-plications. Interestingly, this fact by it-self may have caused a bias in the resultsin favor of the restricted group. Seven

percent of patients in the standard grouprequired repeated surgical interventionsdue to bleeding and the mortality rate ofthis group was 4.7%. We find thesefigures alarming in the setting of elec-tive colorectal surgery, moreover inASA I-II patients. The authors do notstate whether each complication oc-curred in a different patient or could itbe that the same patient that had ananastomotic leakage also suffered fromsepsis and intestinal obstruction. Al-though this is quite possible, it wouldaffect the calculated results for the num-ber of complications per patients.

To conclude, the work presentedhere demonstrated some flaws, whichwill limit the interpretation and applica-tion of these results. Hence, a properlydesigned study is necessary to identifyan adequate fluid resuscitation protocolin postoperative surgical patients.

Yoav Mintz, MDYoram G. Weiss, MD

Avraham I. Rivkind, MD, FACSHadassah University Hospital

Jerusalem, [email protected]

REFERENCES1. Brandstrup B, Tonnesen H, Beier-Holgersen R,

et al. Effects of intravenous fluid restriction onpostoperative complications: comparison oftwo preoperative fluid regimens: a randomizedassessor-blinded multicenter trial. Ann Surg.2003;238:641–648.

2. Finney SJ, Zekveld C, Eia A, et al. Glucosecontrol and mortality in critically ill patients.JAMA. 2003;209:2041–2047.

3. Alderson. Survival: colloids vs. crystalloids forfluid resuscitation in critically ill patients. Co-chrane Database Syst Rev. 2000;2.

4. Cochrane Injuries Group Albumin Reviewers.Human albumin administration in critically illpatients: systematic review of randomized con-trolled trials. BMJ. 1998;317:235.

In Reply:On behalf of the “Danish study

group on perioperative fluid therapy” Ithank Dr. Mintz and colleagues for theircomments on our work, and for givingme this opportunity to report furtherdetails on the trial.

Letters to the Editor Annals of Surgery • Volume 240, Number 2, August 2004

© 2004 Lippincott Williams & Wilkins386

As pointed out by Dr. Mintz andcolleagues, and as reported in the paper,there was no firm protocol for postoper-ative fluid therapy in the standard group.The standard postoperative protocols ofthe centers were to give 1–2 L of intra-venous fluid the rest of the day of surgery,and thereafter supplement oral fluid for atotal amount of 2–3 L daily. I agree thatthis may have caused some degree ofinhomogeneous treatment of the controlgroup and that inhomogeneous treatmentcould have caused a negative result of thetrial due to a small postoperative fluiddifference between the groups. Despite ofthis, however, a marked difference in out-come was found.

Likewise were antibiotic- and an-tithrombosis prophylaxis given to all pa-tients in accordance with the routine ofthe departments. These factors werementioned in the paper because we, likeDr. Mintz and colleagues, find them ofimportance for outcome following sur-gery. Other factors may be of equalimportance: differences in postoperativemobilization regimens, the use of irriga-tion, the use of drains and urinary cath-eters, suture material, anastomosis for-mation, the skills of the surgeon, etc. Asreported in the paper, many confounderswere controlled by exclusion criteria,standardization of treatment and stratifi-cation of randomization. We realized,however, that controlling all known (andespecially unknown) factors of impor-tance for outcome in a multicenter trialwas not possible. Instead we relied onthe benefit of block randomization toensure that each center contributed withan equal number of patients in the 2groups compared. In our opinion, it is ofless importance if differences in varioustreatments and routines exist betweencenters, as long as the centers contrib-utes with an equal number of patients tothe 2 groups compared, where the fluidtreatment is the only difference. Theblock randomization worked very well.

I agree with Dr. Mintz and col-leagues that the major problems withfluid resuscitation occur in ASA-group

III-IV patients. However, the aim of thetrial was to test the hypothesis that stan-dard fluid therapy actually caused car-diopulmonary and tissue-healing com-plications also in patients with nohistory of cardiopulmonary diseases. Asseen from Table 2 in the paper the num-ber of patients with cardiopulmonarydiseases was similar between groups,while the number of patients with post-operative complications was not.

As pointed out by Dr. Mintz andcolleagues the restricted regimen andtype of fluid is not currently practiced(with the exception of the majority ofDanish hospitals and an increasing num-ber of hospitals in the rest of Scandina-via) nor advised. The trial is the first toexamine the paradigm that fluid lossshould be replaced qualitatively andquantitatively but fluid overload (recog-nized as a weight gain) should beavoided. Consequently water lost as per-spiration was replaced with a “waterpreparation” ie, glucose 5%. I disagreein the point of view; the literature dis-courages glucose infusions to normalelective surgical patients. Preoperativeglucose load given intravenously ororally has been shown to reduce thepostoperative insulin resistance1,2 andimprove the muscle strength.3 In theonly randomized trial found testing theeffect of intraoperative glucose admin-istration, the glucose group had the bestoutcome.4 New randomized trials are,however, needed to determine the roleof glucose infusions in major surgery.As pointed out by Dr. Mintz and col-leagues the role of albumin is controver-sial. I am looking forward to the resultsof ongoing trials. One major point must,however, be made regarding albumin.Compared with our restricted regimen,all trials of crystalloids versus colloidsuntil this date have tested the effect offluid overload with crystalloids versusfluid overload with colloids. I am notsurprised if overload with colloids maybe more hazardous for the patients.

Dr. Mintz and colleagues claimthat patients undergoing colorectal sur-gery normally do not require the fluid

volume administered in the trial. Some-times, however, especially in rectal sur-gery large blood losses do occur. Thepatient in the restricted group who re-ceived 8050 mL fluid on the day ofoperation was a 76-year-old woman un-dergoing rectal extirpation who lost ap-proximately 1.5 times her calculatedblood volume. Her weight on the firstpostoperative day was increased by only100g. In my opinion, the investigatinganesthetist did a splendid job in difficultcircumstance.

The patient in the standard groupreceiving 11083 mL was a 75-year-oldmale, undergoing low anterior resectionwho had an initial intraoperative bloodloss of 1400, but continued to bleed post-operatively (1100 mL through drains in afew hours) and underwent reoperation prohemostasis with additional blood loss (andthird space replacements).

Dr. Mintz and colleagues are rightto emphasize that the level of colonic(but not rectal) anastomosis was differ-ent between the groups and may havehad a beneficial influence towards therestricted regimen. As already discussedin the paper, we have no reason to believethat the level of colonic anastomosis influ-ences other complications than anasto-motic leakage. Only one anastomotic leakregistered in this trial affected a colonicanastomosis and this was a leak of thecoecum, all other leaks affected anastomo-sis including the rectum.

Dr. Mintz and colleagues pointsout that multiple complications in onepatient may affect the number of com-plications per patient shown in the pa-per’s Table 4. Table 4 served severalpurposes. First, to define the criteria foracceptance of a complication. Second,to show what complications were regis-tered at all. Third, to illustrate thedistribution of complications into sub-groups. These analyses and the compli-cations included were planned by proto-col before anyone knew the results ofthe trial. Fourth, to illustrate that notonly the number of patients with a com-plication but also the severity and num-ber of complications were marked in-

Annals of Surgery • Volume 240, Number 2, August 2004 Letters to the Editor

© 2004 Lippincott Williams & Wilkins 387

creased in the standard group. In theR-group no patient had multiple compli-cations, as indeed some patients in theS-group. As reported in the paper, theroutine of the department was fol-lowed for patients needing reoperationor intensive care (ie, they were givenstandard fluid replacements). Whileanalyzing these results, it became clearto us that if 1 operation with standardfluid replacement was harmful, then 2operations with an additional standardfluid load in some cases ended in di-saster. Patients treated by the re-stricted regimen during the primaryoperation tolerated a secondary opera-tion with standard fluid replacementmuch better.

I agree with Dr. Mintz and col-leagues in the point of view that theresults of all trials should be reproducedin additional trials. I hope, however, thatthese responses to the raised problemswill convince the readers of the Annalsof Surgery as well as Dr. Mintz andcolleagues that despite a few weak-nesses, this trial is indeed both properlydesigned and accomplished.

Birgitte Brandstrup, MD, PhDGlostrup University Hospital

Glostrup, [email protected]

REFERENCES

1. Ljungqvist O, Thorell A, Gutniak M, et al.Glucose infusion instead of preoperative fast-ing reduces postoperative insulin resistance.J Am Coll Surg. 1994;178:329–336.

2. Nygren J, Soop M, Thorell A, Sree Nair K,Ljungqvist O. Preoperative oral carbohydratesand postoperative insulin resistance. Clin Nutr.1999;18:117–120.

3. Henriksen MG. Effects of preoperative oralcarbohydrates and peptides on postoperativeendocrine response, mobilization, nutrition andmuscle function in abdominal surgery. ActaAnesthesiol Scand. 2003;47:191–199.

4. Cook R, Anderson S, Riseborough M, BloggCE. Intravenous fluid load and recovery. Adouble-blind comparison in gynaecological pa-tients who had day-case laparoscopy. Anaes-thesia. 1990;45:826–830.

Effect of Duodenal-Jejunal Exclusion in aNon-Obese AnimalModel of Type 2Diabetes: A New

Perspective for an OldDisease

To the Editor:I found the article by Rubino and

Marescaux1 very interesting. The authorinvestigated the effect of surgery on type 2diabetes observed in cases of obese pa-tients with type 2 diabetes who underwentRoux-en-Y gastric bypass (GBP) and bil-iopancreatic diversion (BPD).2–6

To determine if long-term controlof blood glucose following surgery wasdue to the treatment of obesity or toalterations in the enteroinsular axis in-duced by duodenal-jejunal exclusion, agastrojejunal bypass (GJB) was performedon non-obese rats with type 2 diabetes.After treatment, reduced fasting glycemiawas observed at an even higher level withrespect to a control group treated withRosiglitazone, with increased glucosetolerance, better insulin sensitivity, andlower levels of FFA and cholesterol. Moreimportantly, blood glucose control wasachieved in non-obese rats without anypostoperative weight loss. This experi-mental model definitively demonstratesthat reduced fasting glycemia and insulinresistance, as well as improved glucosetolerance are attributable to surgery ratherthan solely to weight loss.1

We would like to point out thatthis concept has already been demon-strated in a series published by Noya etal in 1998, presenting a case study of 10moderately obese patients (mean BMIof 33.20 kg/m2) who underwent bilio-pancreatic diversion preserving thestomach and pylorus, a duodenal-jejunalswitch without the restrictive gastricsurgery as that proposed by Rubino inhis paper, to treat hypercholesterolemia,hypertriglyceridemia and type 2 diabe-

tes mellitus. In all treated patients, cho-lesterol and triglyceride levels normal-ized and blood glucose stabilized withinnormal range in 9 patients during thefirst few weeks postoperatively, despitethe fact that no dietary restrictions wereapplied and before a significant weightloss was gained.7

Rubino attributed the result of di-abetes control to duodenal-jejunal ex-clusion, suggesting a potential role ofthe proximal gut in the pathogenesis ofthe disease and putting forward the pos-sibility of alternative therapeutic ap-proaches for the management of type 2diabetes.1 In fact, he has focused atten-tion on glucose-dependent-insulino-tropic peptide (GIP) a secretin producedby the duodenal K-cells that presents amarked decrease in its insulinotropic ef-fect in type 2 diabetic patients.1,8

Though GJB has not produced sig-nificant changes on its secretion, the au-thor proposes that the deficit of the entero-insular system corrected by surgery lies inthe bypassed duodenal-jejunal tract. In-stead, we believe that the physiopathologyof Glucagon-like Peptide 1 (GLP- 1), asecretin produced by the L-cells of theterminal ileum in patients with type 2diabetes mellitus, and its alterations fol-lowing surgery can provide a more likelyexplanation for the resolution of diabetesobserved by Rubino.

GLP-1 is a peptide secreted by theL-cells of the terminal ileum in responseto nutrients and neural stimuli. It exertsa powerful insulinotropic action, the so-called incretin effect, delays gastricemptying, increases satiety and fullness,and has anabolic, glycogenic and lipo-genic actions on liver, fatty and muscletissues.9

In type 2 diabetic patients, the in-cretin effect of GLP-1 is diminished ordisappears entirely, as occurs with GIP,therefore the deficit of entero-insularaxis theorized by Rubino has been al-ready demonstrated.10

Radioimmunoassay of GLP-1 hasshown that this event can be attributed toits reduced secretion in both basal con-ditions and postprandially. Oral glucose

Letters to the Editor Annals of Surgery • Volume 240, Number 2, August 2004

© 2004 Lippincott Williams & Wilkins388

load during a euglycemic hyperinsuline-mic clamp has demonstrated that there isa lower rate of GLP-17-36amide in obesepatients with type 2 diabetes both beforeand after load.11

A reduced GLP-1 response wasalso observed following a mixed meal inobese patients with type 2 diabetes, ascompared with a control group of normo-glycemic obese patients.12,13 The secre-tion of GIP, on the other hand, is normal intype 2 diabetic patients, but its effect islost.10 The reduced incretin effect ofGLP-1 can in fact be attributable to im-paired secretion, whereas with GIP thereduced effect can be attributed to a defectof its receptors and this makes GIP uselessas a hormone for treating type 2 diabetes.8

Moreover, each surgical proce-dure (jejunoileal bypass, GBP, BPD)that produces the early arrival of food atthe terminal ileum triggers the hyperse-cretion of GLP-1 and accompanies theresolution of type 2 diabetes mellitus, asnoted in animal models and in hu-mans.14–18 Therefore, it is our opinionthat GJB has had such an effect ondiabetes, as reported by Rubino, becauseof early arrival of indigested food in theterminal ileum and the consequent stim-ulation of GLP-1 secretion.

In accordance with Mason, newresearch prospects are open for surgicalmethods that can increase GLP-1 secre-tion even in normal-weight or moder-ately obese patients:17 GJB, as proposedby Rubino in animal models and byNoya in humans, may represent one ofthese methods. This becomes particu-larly important due to the fact that phar-macological research is trying to de-velop GLP-1 synthetic analogs with aclinical application for the treatment oftype 2 diabetes mellitus. However, theirtherapeutic utility is still limited by theirshort half-life (1–2 minutes).19

Alberto Patriti, MD,Enrico Facchiano, MD,

Annibale Donini, MDUniversity of Perugia

Perugia, [email protected]

REFERENCES1. Rubino F, Marescaux J. Effect of duodenal-

jejunal exclusion in a non-obese animal modelof type 2 diabetes: a new perspective for anold disease. Ann Surg. 2004;239:1–11.

2. Scopinaro N, Adami GF, Marinari GM, et al.Biliopancreatic diversion. World J Surg.1998;22:936–946.

3. Pories WJ, MacDonald KG Jr., Flickinger EG,et al. Is type II diabetes mellitus (NIDDM) asurgical disease?

4. De Maria EJ, Sugerman HJ, Kellum JM, et al.Result of 281 consecutive total laparoscopicRoux-en-Y gastric bypasses to treat morbidobesity. Ann Surg. 235:640–647.

5. Schauer PR, Burguera B, Ikramuddin S, et al.Effect of laparoscopic Roux-en-Y gastric by-pass on type 2 diabetes mellitus. Ann Surg.2003;238:467–485.

6. Pories WJ, MacDonald KG Jr, Flickinger EG, etal. Is type II diabetes mellitus (NiDDM) a sur-gical disease ? Ann Surg. 1992;215:633–642.

7. Noya G, Cossu ML, Coppola M, et al.Biliopancreatic diversion preserving the stom-ach and pylorus in the treatment of hypercho-lesterolemia and diabetes type II: result in thefirst 10 cases. Obes Surg. 1998;8:67–72.

8. Meier JJ, Nauch MA, Schmidt WE, et al.Gastric inhibitory polypeptide: the neglectedincretin revisited. Regul Pept. 2002;107:1–13.

9. Kieffer TJ, Habener JF. The glucagon-like pep-tides. Endocrine Reviews. 1999;20:876–913.

10. Nauck M, Stockmann F, Ebert R, et al.Reduced incretin effect in type 2 (non-insulin-dependent) diabetes. Diabetologia. 1986;29:46–52.

11. Mannucci E, Ognibene A, Cremasco F, et al.Glucagon-like peptide (GLP)-1 and leptin con-centrations in obese patients with type 2 diabetesmellitus. Diabet Med. 2000;17:713–719.

12. Vilsboll T, Krarup T, Deacon CF, et al.Reduced postprandial concentrations of intactbiologically active glucagon-like peptide 1 intype 2 diabetic patients. Diabetes. 2001;50:609–613.

13. Toft-Nielsen MB, Damholt MB, Madsbad S,et al. Determinants of the impaired secretionof glucagon-like peptide-1 in type 2 diabeticpatients. J Clin Endocrinol Metab. 2001;86:3717–3723.

14. Naslund E, Backman L, Holst JJ, et al.Importance of small bowel peptides for theimproved glucose metabolism 20 years afterjejunoileal bypass for obesity. Obes Surg.1998;8:253–260.

15. Sarson DL, Scopinaro N, Bloom SR. Guthormone changes after jejunoileal (JIB) orbiliopancreatic (BPB) bypass surgery for mor-bid obesity. Int J Obes. 1981;5:471–480.

16. Kellum JM, Kuemmerle JF, O’Dorisio TM, etal. Gastrointestinal hormone responses tomeals before and after gastric bypass andvertical banded gastroplasty. Ann Surg. 1990;211:763–770.

17. Mason EE. Ileal transposition and enteroglu-cagon/GLP-1 in obesity surgery. Obes Surg.1999;9:223–228.

18. Naito H, Matsuno S. Surgical aspect of En-teroinsular Axis after gastrointestinal surgerywith reference to incretin secretion. Pancreas.

1995;16:370–378.19. Holz GG, Chepurny OG. Glucagon-like pep-

tide-1 synthetic analogs: new therapeutic agentsfor use in the treatment of diabetes mellitus.Curr Med Chem. 2003;10:2471–2483.

In Reply:We thank Dr. Patriti and col-

leagues for their interest in our paper.One of their arguments is that the con-cept of a direct antidiabetic effect ofbariatric surgery had already been dem-onstrated by Noya et al with an uncon-trolled case-series of 10 moderatelyobese patients undergoing a stomach-preserving biliopancreatic diversion.1

We are afraid that this claim is notsupported scientifically.

First, a small, uncontrolled case-series type of study is not the properinstrument to demonstrate a direct effectof surgery on type 2 diabetes (T2D) asthere are several possible reasons thatcould justify improved glycemia after abariatric operation. For instance, sincepatients undergoing Roux-en-Y gastricbypass (RYGB) or biliopancreatic di-version (BPD) eat small, rather fluid andlow-caloric meals in the early postoper-ative period, it is admittedly impractica-ble to rule out that the rapid normaliza-tion of plasma glucose and improvedinsulin resistance after these surgeriesbe simply the effect of decreased caloricintake. To rule out this possibility onewould need to do a comparative studywith matched subjects undergoing pair-feeding or a period of strict restriction offood intake, ideally with a random allo-cation. This is a quite difficult study toperform clinically, and has not beendone as of yet. We decided to assess theissue with an animal investigation,which, of course, lend itself better to theset up of these experimental conditions.

Although several independent ob-servations documented rapid remissionof T2D after RYGB and BPD,2–4 allthese studies had not been designed tospecifically test the efficacy of surgeryas a treatment of T2D. The indicationfor surgery was indeed morbid obesity.Hence, the recruitment of patients withdiabetes was unintentional and one

Annals of Surgery • Volume 240, Number 2, August 2004 Letters to the Editor

© 2004 Lippincott Williams & Wilkins 389

could rightfully argue that some of thepatients included in these series mightnot necessarily resemble the typical type2 diabetic population. Furthermore, dueto the strict epidemiologic associationmany experts believe that a causativerole of obesity in type 2 diabetes melli-tus is beyond doubt.5 On the other hand,nonsurgical interventions aimed at fight-ing obesity and hyperlipidemia have awell known beneficial effect on T2D.6

Consequently, even if control of T2Dprecedes significant weight loss wecould not totally exclude that the effectis secondary to the treatment of obesity.Surgical weight loss indeed might justbe a surrogate marker of the improve-ment of other metabolic abnormalitiesrelated to obesity. The report of Noya etal1 did not evidently solve these con-cerns as their observations were madeon an obese population with severe hy-perlipidemia. However, an interestingfindings was that control of T2D oc-curred in 9 out of 10 patients with meanBMI at 33.2 kg/m2, supporting the al-ready plausible hypothesis that an arbi-trary cut-off at BMI 35 is very unlikelyto represent a natural limit for the effec-tiveness of RYGB and BPD on T2D.

By documenting control of T2D ina non-obese animal model of T2D ourstudy rejects the argument that surgicalcontrol of diabetes by duodenal-jejunalbypass is a secondary outcome of thetreatment of obesity.

The second argument of Patriti andcolleagues is that the stomach-preservingmodified BPD performed by Noya et al isthe same as our model of duodenal-jejunalexclusion (DJE). This is also inaccurate asthe difference is quite substantial.

In fact, the hypothesis that DJEcan achieve, “per se,” control of T2D,7

could not be reliably verified by usingstandard RYGB or BPD techniques, asrestriction of the stomach (as in RYGB)and diversion of bile and pancreatic juicesdown to the terminal ileum (as in BPD)would exert confounding interference andmetabolic effects that influence diabeticoutcomes.

We therefore developed a modelthat minimizes influence from other fac-tors. That’s why we not only avoidedmechanical restriction of food intake bypreserving the stomach but we also care-fully set the length of the biliary andalimentary limb of our Roux-en-Y re-construction to minimize the risk of nu-trient malabsorption. In fact, unlike themodified-BPD of Noya et al, our modelleaves most of the intestinal mucosaexposed to the mix of bile and nutrientsand this allows to more specifically ad-dress the role of the proximal bowelbypass in the treatment of T2D.

Although we believe our studydemonstrated a direct effect of DJE onT2D, it does not explain yet what exactlymakes this effect possible. Our study in-deed strengthens the hypothesis that anendocrine effect be involved in the surgi-cal control of T2D, but, it remains unclearwhich hormone response induced by DJEis determinant in the control of T2D. Itmay be either the production of a “protec-tive” factor enhancing insulin sensitivityand/or insulin secretion or the suppressionof a gastrointestinal signal produced in theduodenum-jejunum and causing insulinresistance or strictly involved in its patho-genesis.

As we said in our paper, a possiblecandidate for the first of these 2 hypoth-esis is GLP-1, an incretin hormone thatenhances insulin secretion. However,whereas increased GLP-1 levels havebeen reported after jejuno-ileal bypass,8

more recent studies consistently failed todemonstrate significant GLP-1 changesafter RYGB.9–10 Hence, we would bemore cautious than Patriti and colleaguesin considering the changes in GLP-1 asthe hormonal mechanism by which DJEcontrols T2D; at least until more evidencebecomes available. In contrast, severalstudies consistently showed that glucose-dependent insulinotropic polypeptide(GIP) levels fall shortly after RYGB9–10

and we think this has potential implica-tions in surgical control of T2D. Reducedlevels of GIP may be a downstream effectof one or more other coordinate hormonalchanges that improve insulin sensitivity/

secretion. Indeed, the acute insulin re-sponse to glucose, which is influenced byGIP,11 is characteristically attenuated inT2D patients and seems to be normalizedby RYGB,12 in spite of reduced circulat-ing levels of the hormone. This suggeststhat surgery may reverse a sort of GIP-resistance in T2D.

The demonstration that surgery candirectly influence T2D as opposed to be-ing a secondary effect of the treatment ofobesity is not a mere intellectual exercise;it has, instead, important implications.One is that it implies the new concept of“diabetes surgery” as an independent newsurgical discipline for which surgeonsneed to develop specific knowledge andcompetence. Indeed, clinical studies withdiabetes-specific endpoints are now justi-fied to define whether or not surgical treat-ment of type 2 diabetes should be ex-tended also to moderately obese oroverweight patients as well as which sur-gical technique has the best risk/benefitratio and whether there are specific indi-cations and contraindications for surgicaltreatment of type 2 diabetes.

We are also confident that ourtechnique of DJE may become a valu-able model for diabetes research atlarge. In fact, future investigationsaimed to define its mechanism of actionmight help finding new molecular tar-gets for medical treatment of type 2diabetes and possibly even shed light onthe causes of the disease.

In a time in which the worldwideincreasing incidence of type 2 diabetesposes a very serious threat to health caresystems surgeons can and should maketheir part in the fight against this disease.

Francesco Rubino, MD,Jacques Marescaux, MD, FRCS

University Louis PasteurStrasbourg, France

[email protected]

REFERENCES1. Noya G, Cossu ML, Coppola M, Tonolo G,

Angius MF, Fais E, Ruggiu M. Biliopancreaticdiversion preserving the stomach and pylorus inthe treatment of hypercholesterolemia and dia-betes type II: results in the first 10 cases. ObesSurg. 1998;8:67–72.

Letters to the Editor Annals of Surgery • Volume 240, Number 2, August 2004

© 2004 Lippincott Williams & Wilkins390

2. Pories WJ, Swanson MS, MacDonald KG, etal. Who would have thought it? An operationproves to be the most effective therapy foradult-onset diabetes mellitus. Ann Surg. 1995;222:339–50.

3. Scopinaro N, Adami GF, Marinari GM, et al.Biliopancreatic diversion. World J Surg.1998;22:936–46.

4. Schauer PR, Burguera B, Ikramuddin S,Cottam D, Gourash W, Hamad G, Eid GM,Mattar S, Ramanathan R, Barinas-Mitchel E,Rao RH, Kuller L, Kelley D. Effect of laparo-scopic Roux-en Y gastric bypass on type 2diabetes mellitus. Ann Surg. 2003;238:467–84.

5. Pinkney J, Kerrigan D. Current status of bari-atric surgery in the treatment of type 2 diabe-tes. Obes Rev. 2004;5:69–78.

6. Williamson DF, Thompson TJ, Thun M,Flanders D, Pamuk E, Byers T. Intentionalweight loss and mortality among overweightindividuals with diabetes. Diabetes Care.2000;23:1499–504.

7. Rubino F, Gagner M. Potential of surgery forcuring type 2 diabetes mellitus. Ann Surg.2002;236:554–9.

8. Mason EE. Ileal transposition and enteroglu-cagon/GLP-1 in obesity (and diabetic?) sur-gery. Obes Surg. 1999;9:223–8.

9. Rubino F, Gagner M, Gentileschi P, et al. Theeffect of Roux-en-Y gastric bypass on hor-mones involved in body weight regulation andglucose metabolism. Ann Surg. 2004;in press.

10. Clements RH, Gonzalez QH, Long CI, WittertG, Laws HL. Hormonal changes afterRoux-en Y gastric bypass for morbid obesityand the control of type-II diabetes mellitus.Am Surg. 2004;70:1–4.

11. Lewis JT, Dayanandan B, Habener JF, KiefferTJ. Glucose-dependent insulinotropic polypep-tide confers early phase insulin release to oralglucose in rats: demonstration by a receptorantagonist. Endocrinology. 2000;141:3710–6.

12. Polyzogopoulou EV, Kalfarentzos F, VagenakisAG, Alexandrides TK. Restoration of euglyce-mia and normal acute insulin response to glu-cose in obese subjects with type 2 diabetesfollowing bariatric surgery. Diabetes. 2003;52:1098–103.

LaparoscopicAdjustable Silicone

Gastric Banding VersusVertical BandedGastroplasty inMorbidly Obese

Patients

To the Editor:We read with interest the article by

Morino et al.1 We congratulate the authorson their nicely performed first randomized

trial to compare different laparoscopicbariatric procedures. The study demon-strates that the laparoscopic restrictiveprocedure, either laparoscopic verticalbanded gastroplasty (LVBG) or laparo-scopic adjustable silicone gastric banding(LASGB) are safe, minimally invasiveprocedures. However, we have reserva-tions about their conclusion that LVBG ismore effective than LASGB in terms oflate complications, reoperations, andweight loss. We want to provide 2 com-ments, which might result in this quireconclusion.

The author had a 0% late reopera-tion rate in LVBG which is very un-usual. We started to perform LVBGsince 1998 and had accumulated expe-riences over 600 cases. We had similarresults with the authors that LVBG is asafe and effective mini-invasive bariat-ric operation with only 1% major com-plication rate reported previously.2–4

Our procedure had been observed byMason and regarded as a easy but stan-dard procedure.5The excessive weightloss following surgery in our patients is53.1% at 2 years following surgery,which is similar to the author’s resultsand other studies from European pa-tients.1,6,7 However, LVBG has a majordisadvantage of causing gastrointestinalsymptoms, because patients are unableto eat regular food and the incidence ofvomiting is very high. In our previousstudy, we found that the gastrointestinalquality of life improved significantly at6 months after LVBG but returned topreoperative values at 1 year after surgery,despite an 81% patient satisfaction rate.8

In addition, some patients may regainbody weight gradually after 2 years.Therefore, a reoperation surgery is un-avoidable in patients with VBG either dueto intolerance or inadequate weight loss.9

We had a 7.7% accumulated reoperationrate at 5 years after LVBG. Half of therevision surgeries were changing to gastricbypass and the other half were gastro-gastric bypass with regain of body weight,all by laparoscopic surgery. The 0% latereoperation rate in the authors’ series isunusual. The reasons might be that the

follow-up period is not long enough or thesurgeons tend to neglect the requirementfrom the patients.

On the other hand, the authors hada 24.5% reoperation rate in LASGBgroup and 20% of the bands were re-moved. This result is also unusual for anexperience hand on LASGB. Followingits introduction, the technique ofLASGB underwent several modifica-tions.9,10 After the development of tech-niques of pars flaccida approach forband placement above the bursa omen-talis and gastrogastric suturing knots,the reoperation rates of band are re-ported to decrease to less than 5%. Spe-cifically, O’Brien et al from Australiareported a decreasing reoperation ratefor band slippage from 12.5% to only1% after a learning curve of 350 cases.11

We started to perform the LASGB since2001 after learning the technique fromO’Brien. We had accumulated 81 casesof LASGB until now with a 0% majorcomplication rate, 0% band slippage andonly 2 (1.5%) reoperations are requireduntil now. One patient received port re-fixation for dislocation of the port andthe other required band removal becauseof intolerance. Therefore, the 18% ofband slippage and 20% band removalrate in the authors’ study implicated thattheir techniques are not complete, cor-rect, or are still in the learning curve.

Because the technique performanceis not equal in the authors’ study, theywould have a conclusion that LVBG wassignificantly superior to LASGB in termsof weight loss under the intension-to-treatprinciple. This result can not reflect thedisparity of LVBG and LASGB but onlythe maturity of separate technique of theauthors. In addition, the comparison ofweight reduction between LVBG andLASGB should not be concluded earlierthan 3 years following surgery because thedifferent nature of the devices and dispar-ity of the rate of weight loss. The rate ofweight loss after LVBG is usually rapid inthe first 6 months and then slows downuntil a plateau is reached 1 to 2 years aftersurgery. Patients with LVBG are expectedto regain some weight after 2 to 5 years

Annals of Surgery • Volume 240, Number 2, August 2004 Letters to the Editor

© 2004 Lippincott Williams & Wilkins 391

after surgery.8 On the contrary, the rate ofweight reduction after gastric banding isslow and steady. The expected excessweight loss for banding is 30% to 40% at1 year after surgery. The plateau can bereached 3 years after surgery with a 50%to 60% excess weight loss comparable tothe best of LVBG and further weight losscan be expected even after 5 years.9–12

Our experience with the LSAGB has beenmore similar to the above experience.Therefore, in contrast to the authors’ ex-perience, we have suspend the routineclinical application of the LVBG, and useLSAGB in selected patients in which theadvantages of a less complex and totallyreversible procedure are the principlerequirements determining the surgicaltechnique.

In conclusion, randomized trials tocompare different laparoscopic bariatricsurgeries are essential for the continuingprogress of bariatric surgery. However,because of the technique difficulties andprolonged learning curve of laparo-scopic bariatric surgery, a good qualitycontrol of surgical procedures are man-datory before a conclusion is made.

Wei-Jei Lee MD, PhDWeu Wang, MD

Ming-Te Huang, MDEn-Chu Kong Hospital,

Taipei Hsien, [email protected]

REFERENCES1. Morino M, Toppino M, Bonnet G, et al.

Laparoscopic adjustable silicone gastric band-ing versus vertical banded gastroplasty inmorbidly obese patients: a prospective ran-domized controlled clinical trial. Ann Surg.2003;238:835–842.

2. Lee WJ, Lai IR, Huang MT, et al. Laparoscopicversus open vertical banded gastroplasty for thetreatment of morbid obesity. Surg Lapar, Endo& Percut Tech. 2001;11:9–13.

3. Lee WJ, Yu PJ, Wang W, et al. Gastrointestinalquality of life following laparoscopic verticalbanded gastroplasty. Obes Surg. 2002;12:819–824.

4. Lee WJ, Wang W, Chen TC, et al. Clinicalsignificance of central obesity in laparoscopicbariatric surgery. Obes Surg. 2003;13:921–925.

5. Nguyen N, Goldman C, Rosenquist J, et al.Laparoscopic versus open gastric bypass: arandomized study of outcomes, quality of life,and costs. Ann Surg 2001;234:279–291.

Azagra JS, Goergen M, Ansay J, et al. Lapa-roscopic gastric reduction surgery: prelimi-nary results of a randomized, prospective trialof laparoscopic vs open vertical banded gas-troplasty. Surg Endosc. 1999;13:555–558.

6. Lonroth H, Dalenback J, Haglind E, et al.Vertical banded gastroplasty by laparoscopictechnique in the treatment of morbid obesity.Surg Laparos Endosc. 1996;6:102–106.

7. Nuslund E, Backman L, Granstrom L, et al.Seven years results of vertical banded gastro-plasty for morbid obesity. Eur J Surg. 1997;163:281–286.

8. Gentileschi P, Kini S, Catarci M, et al.Evidence-based medicine: open and laparo-scopic bariatric surgery. Surg Endosc. 2002;16:736–44.

9. Favretti F, Cadiere GB, Segato, et al. Laparo-scopic adjustable gastric bandin (Lap-Band):how to avoid complications. Obes Surg. 1997;7:352–358.

10. Belachew M, Legrand M, Vincent V, et al.Laparoscopic adjustable gastric banding.World J Surg. 1998;22:955–963.

11. O’Brien PE. Brown WA, Smith A, et al.Prospective study of a laparscopically placed,adjustable gastric band in the treatment ofmorbid obesity. Br J Surg. 1999;86:113–118.

12. Zinzindohouse F, Chevallier JM, Douard R, etal. Laparoscopic gastric banding: a minimallyinvasive surgical treatment for morbid obe-sity. Ann Surg. 2003;237:1–9.

LaparoscopicAdjustable Silicone

Gastric Banding VersusVertical BandedGastroplasty inMorbidly Obese

Patients

To the Editor,I read with interest the random-

ized trial published by Dr. Morino andcolleagues1 comparing the laparoscopicadjustable silicone banding (LASGB)with vertical banded gastroplasty inobese patients. I know the difficulties toconduct randomized trials in surgeryand congratulate Dr. Morino and col-leagues for their timely work. The arti-cle typically illustrates the problem ofevaluating an emerging and evolving sur-gical technique, and raises the question ofwhen a randomized trial should be con-ducted in the life of a procedure. TheItalian study began in February 1999, just

before the publication of another Germanrandomized trial suggesting that the eval-uated LASGB procedure to be no longerconsidered as a standard.2

According to the authors, the Lap-Band (Bioenterics, Carpinteria, CA) ispositioned close to the gastric wall. Thisprocedure is now out of date. The cur-rent “standard” is to use the “pars flac-cida route.” In 2001, a randomized trial2

suggested the superiority of this tech-nique over that used by Dr. Morino andcolleagues. Afterward most bariatricsurgeons adopted this technique and ob-served a significant falling of the rate ofpouch dilatation and slippage followingLASGB.3 Thus, the high morbidity ratereported in this trial is, in my opinion,related not to the LASGB itself but to atechnique currently abandoned. Thisfact limits the external validity of thistrial and should bring us to interpret theresults with caution. One can hypothe-size that the “pars flaccida route” couldnot involve such a high morbidity rate.To illustrate this hypothesis, 2 randomizedtrials from Sweden have shown that band-ing carries a smaller risk of reoperationthan vertical banded gastroplasty4 or hassimilar outcomes5 which is also in accor-dance with the conclusions of the reviewpublished by Clegg and colleagues assess-ing the bariatric surgery and invalidatingany advantage of vertical gastroplastyover gastric banding.6

If one assumes that the resultswere reported as “means” and on inten-tion to treat basis, one can consider thatthe poor “mean” results after LASGB interms of weight loss could be due toremoval of 8 bands (nearly 1 patient outof 5). It is obvious that the more bandsare removed (because of food intoler-ance or pouch dilatation or slippage) themore “mean” residual weight excess isimportant. What about the successfulLASGB procedures? In the Swedish tri-al,4 weight loss was greater after band-ing than after vertical gastroplasty.

In conclusion, the trial publishedby Dr. Morino and colleagues did notfavor the LapBand, but it involved a pro-cedure currently abandoned. This consti-

Letters to the Editor Annals of Surgery • Volume 240, Number 2, August 2004

© 2004 Lippincott Williams & Wilkins392

tutes a major flaw, which explains thepoor results after banding and hampers theexternal validity of the trial. In the future,a properly designed trial should evaluaterecognized standard procedures.

K. Slim MD, FACSHotel Dieu

Clermont-Ferrand, [email protected]

REFERENCES1. Morino M, Toppino M, Bonnet G, et al.

Laparoscopic adjustable silicone gastric band-ing versus vertical banded gastroplasty in mor-bidly obese patients. A prospective randomizedcontrolled clinical trial. Ann Surg. 2003;238:835–842.

2. Weiner R, Bockhorn H, Rosenthal R, et al. Aprospective randomized trial of different lapa-roscopic gastric banding techniques for morbidobesity. Surg Endosc. 2001;15:63–68.

3. Dargent J. Pouch dilatation and slippage afteradjustable gastric banding: is it still an issue?Obes Surg. 2003;13:111–115.

4. Nilsell K, Thorne A, Sjostedt S, et al.Prospective randomised comparison of adjust-able gastric banding and vertical banded gas-troplasty for morbid obesity. Eur J Surg. 2001;167:504–509.

5. Lundell L, Ruth M, Olbe L. Vertical bandedgastroplasty or gastric banding for morbid obe-sity: effects on gastro-oesophageal reflux. EurJ Surg. 1997;163:525–531.

6. Clegg A, Colquitt J, Sidhu M, et al. Clinicaland cost effectiveness of surgery for morbidobesity: a systematic review and economicevaluation. Int J Obes Relat Metab Disord.2003;27:1167–1177.

Reply:We thank Dr. Lee and Dr. Slim for

their comments and welcome the oppor-tunity to respond. In his letter, Dr. Leestates that a 0% late reoperation rate afterlaparoscopic vertical banded gastroplasty(LVBG) is unusual compared with hispersonal experience of 7.7% reoperationrate at 5 years due to food intolerance orinadequate weight loss. In our RCT and inour clinical practice we select very care-fully patients to be submitted to restrictivebariatric surgery considering binge eaters,sweet eaters and patients with a BMI � 50as controindications to LVBG. This clini-cal attitude probably explains the differ-ence in reoperation rates after LVBG be-tween Dr. Lee’s series and our series:inadequate weight loss after restrictivebariatric surgery is common in sweet eat-

ers and in superobese patients,1–3 whilefood intolerance is frequent in binge eat-ers.2,4 By carefully preoperatively select-ing patients many series present low reop-eration rates after LVBG: we recentlypublished a 2.2% reoperation rate at 5years,3 the Italian national Registry forBariatric Surgery presents a 1.8% reopera-tion rate at 6 years.5

Dr. Slim states that “two random-ized trials from Sweden have shown thatbanding carries a smaller risk of reop-eration than vertical banded gastro-plasty6 or has similar outcomes.”7 In ourarticle we already stated that these 2trials are at present of limited interest asthey compare open bariatric procedures;furthermore Nilsell still uses the originalMason’s vertical banded gastroplastyand Lundell uses a non-adjustable GoreTex band: both techniques are not per-formed in laparoscopic surgery.

Both Drs. Lee and Slim are sur-prised by the high reoperation rate afterlaparoscopic adjustable silicone gastricbanding (LASGB) in our series. Further-more, Dr Slim states that the techniquewe used is at present abandoned becauseof the high late complication rates, whileDr. Lee states that the technique we usedhas in his hands much better results. Thedisagreement of these 2 authors on theresults of LASGB using the perigastricroute is in itself an answer: slippagerates using the perigastric route in theliterature8–11 varies between 5% and21% (ours is 18%), the perigastric routeis not at all abandoned as it is demon-strated by Dr. Lee’s experience and bynumberless recent publications;8,9,12 infact the perigastric route is still the mostcommon technique of LASGB in Italy5

and we suspect all over the world.Furthermore, Dr. Slim evocates

the high percentage of band removal asan explication to the poor results interms of weight loss in the LASGBgroup. Eight patients had a band re-moved in our series as a consequence ofslippage and/or intolerance with dyspha-gia and vomiting; no difference in termsof mean BMI and EWL at 3 years werefound comparing these patients to the

global LASGB group: BMI 36.2 versus35.9, EWL 37% versus 39%.

Finally, both authors state that along learning curve is requested forLASGB. All LASGB in our series wereplaced by the same surgeon (MM) with aprevious experience of more than 5000laparoscopic procedures, 300 laparoscopicbariatric procedures and 40 LASGB; if alonger learning curve is required very fewsurgeons will be able to complete it.

Mario Morino, MDMauro Toppino, MD

University of TurinTurin, Italy

[email protected]

REFERENCES1. Sugerman HJ, Starkey JV, Birkenauer R. A

randomized prospective trial of gastric bypassversus vertical banded gastroplasty for morbidobesity and their effects on sweet versus nonsweet eaters. Ann Surg. 1987;205:613–624.

2. Mac Lean LD, Rhode BM, Sampalis J, et al.Results of the surgical treatment of morbidobesity. Am J Surg. 1993;165:155–160.

3. Morino M, Toppino M, Bonnet G, et al.Laparoscopic vertical banded gastroplasty formorbid obesity: assessment of efficacy. SurgEndosc. 2002;16:1566–1572.

4. Baltasar A, Bou R, Arlandis F, et al. Verticalbanded gastroplasty at more than 5 years.Obes Surg. 1998;8:29–34.

5. Toppino M, Morino F, Morino M, et al. ItalianRegistry for bariatric surgery: data up to 6years. Obes Surg. 2003;13:607.

6. Nilsell K, Thorne A, Sjostedt S, et al.Prospective randomised comparison of adjust-able gastric banding and vertical banded gas-troplasty for morbid obesity. Eur J Surg.2001;167:504–509.

7. Lundell L, Ruth M, Olbe L. Vertical bandedgastroplasty or gastric banding for morbidobesity: effects on gastro-oesophageal reflux.Eur J Surg. 1997;163:525–531.

8. Favretti F, Cadiere GB, Segato G, et al.Laparoscopic banding: selection and techniquein 830 patients. Obes Surg. 2002;12:385–390.

9. O’Brien PE, Dixon JB, Brown WA, et al. Thelaparoscopic adjustable gastric band (LapBand) a prospective study of medium termeffect on weight, health and quality of life.Obes Surg. 2002;12:652–660.

10. Belachew M, Legrand M, Vincent V, et al.Laparoscopic adjustable gastric banding.World J Surg. 1998;22:955–963.

11. Allen JW, Coleman MG, Fielding GA. Lessonslearned from laparoscopic gastric banding formorbid obesity. Am J Surg. 2001;182:10–14.

12. Belachew M, Belva PH, Desiave C, et al.Long term results of laparoscopic adjustablegastric banding for the treatment of morbidobesity. Obes Surg. 2002;12:564–568.

Annals of Surgery • Volume 240, Number 2, August 2004 Letters to the Editor

© 2004 Lippincott Williams & Wilkins 393