ARTICLE IN PRESS

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/suronc

Surgical Oncology (2007) 16, S105–S108

0960-7404/$ - see frodoi:10.1016/j.suronc

�Corresponding au

E-mail address: g

REVIEW

Protective stoma in anterior resection of the rectum:When, how and why?

Giuseppe Pappalardo�, Domenico Spoletini, Delia Proposito, Fabio Giorgiano,Anna Maria Conte, Fabrizio Maria Frattaroli

Department of General Surgery, Surgical Specialties and Organ Transplantation ‘‘P.Stefanini’’,Division of General Surgery ‘‘I’’, University of Rome ‘‘La Sapienza’’, Italy

KEYWORDSRectal cancer;Anterior resection;Protective stoma;Colostomy;Ileostomy

nt matter & 2007.2007.10.032

thor. Tel.: +39 033

iuseppe.pappalard

SummaryBackground: The use of protective stoma in anterior resection (AR) is controversial.Neoadjuvant therapy, TME and laparoscopy seem to increase the rate of anastomoticdehiscences (a.d.).Patients and methods: In a prospective study, 219 patients were submitted to electiveopen AR (109 patients), open AR+TME nerve-sparing (110 patients), 35 of which hadintrasphinteric anastomosis. Fifty-five patients were treated by neoadjuvant therapy.Primary stoma was not performed.Results: We had 15 (6.8%) a.d.: 5 (2.3%) major and 10 (4.4%) minor. In the five major a.d.an immediate colostomy was performed with one death. In the 10 minor the a.d. was curedconservatively.Conclusions: A protective stoma is necessary in less than 10% of the patients treated withAR, so avoiding further surgery, mortality, morbidity and higher medical costs in mostpatients.& 2007 Elsevier Ltd. All rights reserved.

Contents

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S106Patients and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S106Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S106Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S107Conflict of Interest Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S108References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S108

Elsevier Ltd. All rights reserved.

8 7248316; fax: +39 06 49970853.

o@uniroma1.it (G. Pappalardo).

ARTICLE IN PRESS

G. Pappalardo et al.S106

Introduction

The use of a stoma to protect colorectal anastomosis hasbeen a controversial issue for decades. Despite the largebody of experimental and clinical data, there is no concreteevidence that a stoma can reduce the incidence ofanastomotic leaks [1–3]. The widespread use in clinicalpractice of the ‘‘total mesorectal excision’’ (TME) proposedby Heald and Ryall [4] has shown that this procedure doesreduce local recurrences but may increase the incidence ofanastomotic dehiscences (a.d.). Consequently, most authorsbelieve that TME-treated patients [3,5,6] require a protec-tive primary stoma.

As the frequent use of neoadjuvant therapy seems toincrease the risk of anastomotic leaks after anteriorresection (AR) of the rectum, the use of a primary stomais recommended [7]. Laparoscopy in resective surgeryof rectal cancer also leans towards the use of a primarystoma [8,9].

However, why and when a stoma should be performed,and which stoma (ileo- or colostomy) is preferable, arequestions that remain unanswered. Authors in favor ofthe routine use of a primary stoma claim that it reduces theincidence of anastomotic complications [3,10], whereasthose in favor of the selective use of a primary stomareserve it for high-risk anastomoses [1,11]. As regards thechoice of stoma, an ileostomy is preferred by the majority ofauthors who perform it at the time of the AR [6,12,13],though some prefer a colostomy [3].

Patients and methods

In order to assess the need for a protective stoma in AR ofthe rectum for cancer, we designed a prospective study inwhich the stoma was not used.

The study started in January 1989 and included patientstreated until the end of May 2007. The inclusion criteriawere: all patients with rectal cancer who underwentelective AR, including those treated with TME (from 1997),neoadjuvant therapy and low or intrasphincteric anastomo-sis. Exclusion criteria were: incomplete rings in the stapler,a positive hydropneumatic test and urgent surgery.

The traditional surgical procedure was used, with amedian laparotomy, complete mobilization of the left colicflexure, ligature 2 cm above the origin of the inferiormesenteric artery or after the origin of the left colic artery,and ligature of the inferior mesenteric vein under theinferior border of the pancreas. We used a mechanicaltermino-terminal anastomosis in most cases, reservingdouble mechanical anastomosis (Knight–Griffen) for ultra-low anastomoses. We positioned a retro-anastomotic drai-nage tube in all the patients; the tube was mobilizedimmediately after canalization and removed, in cases inwhich anastomotic complications were absent, 1 or 2 dayslater. We performed a reperitonization, with extraperitoni-zation of the anastomosis.

All the operations were carried out by two surgeons(G.P. and F.M.F.) who had, before starting the study, alreadyperformed over 50 rectal cancer surgical procedures.

Since 1997, all the patients with cancer of the middle andlower thirds of the rectum as well as those with advanced

cancer (T3–T4) of the upper third have been treated withnerve-sparing TME.

A fistula or anastomotic dehiscence was suspected whenpelvic and/or peritoneal pain, fever, leucocytosis andfecaloid liquid in the drainage were present. An anastomoticleak was confirmed by means of angio-CT and/or endoscopy.

We defined a major dehiscence when signs of peritonealreaction and sepsis were present, regardless of the diameterof the fistula; when diagnosed, a transverse colostomy wasimmediately performed. Clinically less serious cases weredefined as minor dehiscences, in which a ‘‘wait and see’’strategy was used.

The aims of this study were to assess:

1.

How many patients effectively required a secondarystoma.

2.

The morbidity and mortality rates in patients treatedwith a secondary stoma operation compared with thosefrom the literature regarding patients treated with aprimary stoma.

Out of 243 patients who underwent resection for cancer ofthe rectum between January 1989 and May 2007, 224patients (92%) fulfilled the inclusion criteria of our protocol.

The patients enrolled were 115 males and 109 females,aged between 34 and 88 years (average age: 64.4 years). In72 patients (32.1%), the tumor was located in the upperthird of the rectum, in 97 (43.4%) in the middle third and in55 (24.5%) in the lower third. One hundred and nine patients(48.8%) were treated with traditional AR, 110 (49%) with ARand nerve-sparing TME, and the remaining 5 (2.2%) withabdomino-perineal amputation. An intrasphincteric anasto-mosis was performed in 35 patients (16%). Fifty-sevenpatients (25%) were treated by means of neoadjuvanttherapy.

Results

The pathological examination revealed 5 Dukes A, 108 DukesB, 96 Dukes C and 15 Dukes D stages.

The free distal margin on the fresh specimen was equal toor more than 20mm, while in cases of double mechanicalanastomosis, the distal clearance margin had to be at least10mm.

We had 15 (6.8%) anastomotic leaks. Five of these (2.2%)were defined as major dehiscences: two underwent TME(2/110 ¼ 1.8%) with a double mechanical intrasphinctericanastomosis (2/35 ¼ 5.7%); two received neoadjuvanttherapy (2/57 ¼ 3.5%) and one (1/109 ¼ 0.9%) wastreated with traditional AR. All five patients underwent aprotective colostomy within hours of the onset of clinicalsymptoms. We did not expose the anastomosis in these fivepatients as this would have required the opening of thepelvic peritoneum and, consequently, increased peritonealcontamination.

Four patients were discharged within 15 days of thecolostomy, after removal of the pelvic drainage andinterruption of the antibiotic therapy. After checking thatthe anastomosis had healed by means of enema, thecolostomy was closed within 3 months in all four of thesepatients.

ARTICLE IN PRESS

Protective stoma in anterior resection of the rectum: When, how and why? S107

One patient (1/219 ¼ 0.4%) died on the 26th post-operative day owing to respiratory complications andcardio-circulatory failure. This patient was 82 years old,ASA III, diabetic and had Dukes stage C cancer.

The 10 minor dehiscences (4.4%) were treated conserva-tively with antibiotics and enteral feeding.

This therapy was administered using an out-patientregimen. Three of these patients (30%) required percuta-neous CT-guided drainage for a pelvic abscess.

Discussion

The use of a primary stoma in AR for cancer of the rectumvaries in the literature from 0% to 100%. There is no evidencethat a primary stoma reduces the rate of anastomoticdehiscence. Wong and Eu [1], in a comparative prospectivestudy on 1078 patients, did not find any difference inthe dehiscence rate between the 742 patients routinelytreated with a primary stoma and the 324 patients in whom asecondary stoma was performed only in case of dehiscence(3.8% versus 4%, respectively). Given the importance of‘‘evidence-based’’ findings in clinical practice, these datacannot be ignored.

The presence of a stoma does, however, make the clinicalimpact of an anastomotic dehiscence less serious.

The questions which remain unanswered are:

1.

What are the mortality and morbidity rates for asecondary stoma performed for a major anastomoticdehiscence?

2.

In what proportion of patients is a protective stoma reallynecessary?

The mortality rate for anastomotic dehiscences without aprimary stoma appears to be unrelated to the presence ofa stoma, but exclusively due to the delay in performing asecondary stoma. Indeed, there is no significant differencein the mortality rate between patients with and thosewithout a stoma: respectively 0.9% and 2%, according toGastinger et al. [3], and 0% for secondary ileostomypromptly performed in case of an anastomotic leak by Traciet al. [2].

In our 219 patients, we only had one death aftersecondary colostomy, which corresponds to 0.4% of thetotal AR patient series and 6.6% of those with anastomoticcomplications. Moreover, the patient was an elderly manwith significant comorbidity and Dukes C cancer. Thismortality rate is not higher than the average mortality ratefor AR with a primary stoma performed routinely in high-riskpatients [1,3].

The data in the literature cannot yet be considereddefinitive. Platell et al. [10] state that ‘‘more than 90% oftheir patients treated with loop ileostomy appeared to havederived no benefit from their de-functioning stoma’’.

A critical analysis of the papers that report a reducedincidence of anastomotic complications following theroutine use of a primary stoma [3,5,10] reveals that patientswere not randomized, the analysis were retrospective andthe groups compared were not homogeneous.

The authors who do not use a primary stoma [1,14] reportthat a secondary stoma was required in 3–7% of the patients,

and all agree that a secondary stoma does not increase themortality rate.

In our experience, the overall dehiscence rate was 6.8%,though a secondary colostomy was necessary in only 2.2% ofthe patients; this means that the stoma would have beenunnecessary in 97.8% of the patients.

Those surgeons who selectively use a stoma in high-riskpatients do not agree on which factors should be defined ashigh risk, with the exception of ASA 3 and 4, immunode-pressive status, urgent surgery, incomplete rings and apositive hydropneumatic test. Indeed, TME, the use ofsteroids, neoadjuvant therapy, major intra-operative he-morrhage and intrasphincteric anastomosis are not unan-imously accepted as risk factors [11,15,16].

We believe that reperitonization helps prevent massiveperitoneal contamination in case of anastomotic dehiscenceas well as pelvic adherence, and reduces small boweldamage in patients who undergo post-operative radio-therapy. Therefore, reperitonization, pelvic drainage andstrict patient monitoring until complete canalization allowboth early diagnosis of anastomotic leaks and promptsurgery to implant a secondary stoma without exposingpatients to excessive risks. Gastinger et al. [3] reportthat 10.1% of their patients with an anastomotic leakrequired a secondary stoma. We should bear in mindthat the dehiscence rate in AR is never higher than 10% inqualified series [1,2,10], and that most of these cases can betreated conservatively. In our experience, 10 out of 15anastomotic dehiscences (66.6%) were successfully treatedconservatively.

What about its use in laparoscopic surgery, whose ‘‘pillar’’is to be a minimally invasive procedure? We must stress thata stoma can cause morbidity, and that its closure is not aminor procedure, it being burdened by morbidity (15–30%)and even, surprisingly, by mortality [3,15,17]. Pokorny et al.[15] report mortality and morbidity rates, respectively, of2% and 15% for the closure of the ileostomy and of 3% and19% for colostomy. Lastly, we should not forget that a stomamight predispose patients to anastomotic stenosis, particu-larly to mechanical anastomosis, presumably owing to thetemporary loss of the plastic function of fecal transitthrough the anastomosis.

Moreover, ‘‘temporary’’ stomas may become definitive,with rates of 12.2% [3], 19.2% [17] and 32% [7] in differentstudies.

As regards the choice of either an ileostomy or colostomyas a primary derivation, while most authors [17–21] preferthe former, some [3] still prefer the latter. After comparingthe two stomas in a meta-analysis, Guenaga et al. [22]conclude that it is not yet possible to determine whichprocedure is preferable. Ileostomy is burdened by a higherincidence of not programmed admissions before its closure,and by more complicated management [23,24], whilecolostomy is burdened by a higher incidence of infectionsof the wound and incisional hernias after closure [3,15,16].There are no comparative data on the rate of severe post-operative complications (stenosis, fistula) after the closureof an ileostomy and that of a colostomy [22].

We prefer a colostomy as a secondary stoma because theentity of peritoneal contamination, in case of a majordehiscence, is reduced by excluding over half of the residualcolon.

ARTICLE IN PRESS

G. Pappalardo et al.S108

In conclusion, on the basis of both our experience andthat of other authors, we believe that an accuratepreparation of the patient, including the choice of surgicaltiming in the case of neoadjuvant therapy and a correctsurgical procedure, obviates the need for a protective stomain most patients with rectal cancer, including those patientstreated with TME and ultra-low anastomosis. By avoidingthis procedure, further surgery, as unjustified morbidity,mortality and higher medical costs are avoided, and, lastbut not least, the patient’s quality of life also improves. Theuse of primary stoma should be reserved for patients whorequire urgent surgery, who have incomplete rings and apositive hydropneumatic test, as well as for very high-riskpatients in whom the clinical impact of a major dehiscencemay be considerable.

Conflict of Interest Statement

The Authors do not have any potential or actual personal,political, or financial interest in the material, information,or techniques described in the paper.

References

[1] Wong N, Eu K. A defunctioning ileostomy does not preventclinical anastomotic leak after a low anterior resection: aprospective, comparative study. Diseases of the Colon andRectum 2005;48(11):2076–8.

[2] Traci L, Hedrick TL, Sawyer RG, Foley EF, Friel CM. Anastomoticleak and the loop ileostomy: friend or foe? Diseases of theColon and Rectum 2006;49(8):1167–76.

[3] Gastinger I, Marusch F, Steinert R, Wolff S, Koeckerling F,Lippert H. Protective defunctioning stoma in low anteriorresection for rectal carcinoma. British Journal of Surgery 2005;92(9):1137–42.

[4] Heald R, Ryall R. Recurrence and survival after totalmesorectal excision for rectal cancer. Lancet 1986;1:1479.

[5] Den Dulk M, Smit M, Koen CM, Van de Velde CJ. A multivariateanalysis of limiting factors for stoma reversal in patientswith rectal cancer entered into the total mesorectal excision(TME) trial: a retrospective study. Lancet Oncology 2007;8(4):297–303.

[6] Slim K, Sastre B. Should a diverting stoma be associated torectal resection for carcinoma? Annales de Chirurgie 2003;128:256–7.

[7] Rodriguez-Ramirez SE, Uribe A, Ruiz-Garcia EB, Labastida S,Luna-Perez P. Risk factors for anastomotic leakage afterpreoperative chemoradiation therapy and low anterior resec-tion with total mesorectal excision for locally advanced rectalcancer. Revista de investigacion clinica 2006;58(3):204–10.

[8] Lelong B, Bege T, Esterni B, Guiramand J, Turrini O, MoutardierV, et al. Short-term outcome after laparoscopic or openrestorative mesorectal excision for rectal cancer: a compara-tive cohort study. Diseases of the Colon and Rectum 2006;50(2):176–83.

[9] Law WL, Lee YM, Choi HK, Seto CL, Ho JWC. Laparoscopic andopen anterior resection for upper and mid rectal cancer: anevaluation of outcomes. Diseases of the Colon and Rectum 49(8): 1108–1115.

[10] Platell C, Barwood N, Makin G. Clinical utility of a de-functioning loop ileostomy. ANZ Journal of Surgery 2005;75(3):147–51.

[11] Bax T, McNevin S. The value of diverting loop ileostomy on thehigh-risk colon and rectal anastomosis. The American Journalof Surgery 2007;193:585–8.

[12] Gooszen AW, Geelkerken RH, Hermans J, Lagaay MB, GooszenHG. Temporary decompression after colorectal surgery: rando-mized comparison of loop ileostomy and loop colostomy. BritishJournal of Surgery 1998;85:76–9.

[13] O’Leary DP, Fide CJ, Foy C, Lucarotti ME. Quality of life afterlow anterior resection with total mesorectal excision andtemporary loop ileostomy for rectal carcinoma. British Journalof Surgery 2001;88:1216–20.

[14] Eckmann C, Kujat P, Schiedeck THK, Shekarriz H, Bruch HP.Anastomotic leakage following low anterior resection: resultsof a standardized diagnostic and therapeutic approach.International Journal of Clinical Practice 2004;19:128–33.

[15] Pokorny H, Herkner H, Jakesz R, Herbst F. Predictor forcomplications after loop stoma closure in patients with rectalcancer. World Journal of Surgery 2006;30(8):1488–93.

[16] Tilney S, Sains P, Lovegrove E, Tekkis P. Comparison ofoutcomes following ileostomy versus colostomy for defunction-ing colorectal anastomosis. World Journal of Surgery 2007;31:1142–51.

[17] Rullier E, Le Toux N, Laurent C. Loop ileostomy versus loopcolostomy for defunctioning low anastomosis during rectalcancer surgery. World Journal of Surgery 2001;25:274–8.

[18] Poon TP, Chu KW, Ho WC. Prospective evaluation of selectivedefunctioning stoma for low anterior resection withtotal mesorectal excision. World Journal of Surgery 1999;23:463–78.

[19] Edward DP, Leppington-Clarke A, Sexton R, Heald RJ, Moran BJ.Stoma-related complications are more frequent after trans-verse colostomy than loop ileostomy: a prospective rando-mized clinical trial. British Journal of Surgery 2001;88:360–3.

[20] O’Toole GC, Hyland MP, Grant DC, Barry MK. Defunctioning loopileostomy: a protective audit. American College of Surgeons2001;188(1):6–9.

[21] Kalady MF, Ryan CF, Klein S. Loop ileostomy closure at anambulatory surgery facility. Diseases of the Colon and Rectum2003;46(4):486–91.

[22] Guenaga KF, Lustosa S, Saad S, Saconato H, Matos D. Ileostomyor colostomy for temporary decompression of colorectalanastomosis. Cochrane database of systematic reviews, 2007,Issue 1, Art no. CD004647.

[23] O’Leary DP, Fide CJ, Foy C, Lucarotti ME. Quality of life afterlow anterior resection with total mesorectal excision andtemporary loop ileostomy for rectal carcinioma. British Journalof Surgery 2001;88:1216–20.

[24] Tomonori H, Haruhiko N, Tatsuya O. S-III-02 construction andmanagement of loop ileostomy. Wound Repair and Regenera-tion 2004;12(1):A3.