ORIGINAL ARTICLE

Management and outcome of colorectal anastomotic leaks

Michael Thornton & Heman Joshi &Chandrakumar Vimalachandran & Richard Heath &

Paul Carter & Ufuk Gur & Paul Rooney

# Springer-Verlag 2010

AbstractPurpose Anastomotic leak is a devastating complication ofan intestinal anastomosis. Optimal management and outcomeis not routinely described, and much of our knowledge reliesupon historical data. We wished to examine the managementand outcome of anastomotic leaks on a colorectal surgery unitin the twenty-first century.Method A retrospective audit of all patients who had acolorectal anastomotic leak between January 2002 andDecember 2008 in a large university teaching hospital. Datacollected included patient characteristics, primary diagnosis,mode of diagnosis and time to diagnosis of anastomotic leak,inpatient management, morbidity and mortality, permanentstoma rate, use of hospital resources.Results Thirty patients (16 male, 14 female), with a medianage of 60 years (range 25–84 years), had an anastomotic leak.The median time to presentation of clinically suspected leakswas 12 days (range 3–56 days). Fourteen patients requiredreoperation, with ten needing the anastomosis take down.Average hospital stay was 40 days. The permanent stoma ratefollowing a rectal anastomotic leak was 27% and 57.1% froma colonic leak. Overall mortality in this series was 27%.Mortality was higher after leak from a colonic anastomosisthan after leak from a rectal anastomosis (43.8% vs. 7.1%,respectively).Conclusions Anastomotic leaks are not detected until latein the post-operative period and are associated with a highmortality. Demand on hospital resources is high. In this

series, patients who leaked after a colonic anastomosis hada higher mortality and permanent stoma rate than after leaksfrom a rectal anastomosis.

Keywords Anastomotic leak . Colorectal surgery .

Mortality . Stoma

Introduction

A leak from an intestinal anastomosis is the complicationmost feared by the colorectal surgeon. Reported colonicanastomosis leak rates range from 1.5% up to 16% withmortality rates typically quoted between 10% and 20% [1–5].It is generally accepted that rates will vary between surgeons,from unit to unit, and that anastomotic leak is more commonafter anterior resection than colonic resection [6]. In additionto a raised mortality in the early post-leak period, patientsundergoing curative colorectal resection have been demon-strated to have poorer long-term functional results, increasedlocal recurrence rates and reduced 5-year cancer-specificsurvival after anastomotic leak [7–10].

Focus is often placed on examining aetiology of leaksand leak rates, with less information on outcomes andmanagement following a leak. In this era of subspecialisa-tion of colorectal surgery, advances in critical care,antibiotic therapy, enhanced recovery protocols, focusedtraining, modern stapling devices and superior radiologicalimaging and interventional techniques, much of thepublished data may be less relevant to the twenty-firstcentury colorectal surgeon.

The largest published modern series by Francone et al.,reviewed leaks from 79 patients from 2001 to 2007 at a singleunit. The leak rate for this period was 3%, with mortality of10.1% and a permanent stoma rate of 71.8% [11].

M. Thornton (*) :H. Joshi :C. Vimalachandran : R. Heath :P. Carter :U. Gur : P. RooneyDepartment of Colorectal Surgery,Royal Liverpool University Hospital,Prescot Street,Liverpool L7 8XP, UKe-mail: mpt@liv.ac.uk

DOI 10.1007/s00384-010-1094-3Int J Colorectal Dis (2011) 26:313–320

Accepted: 8 November 2010 /Published online: 25 November 2010

In this study, we examine the management and outcomesfollowing colorectal anastomotic leaks on a moderncolorectal unit in a large central teaching hospital, with aparticular emphasis on time of diagnosis, mortality andpermanent stoma rate.

Method

A retrospective audit was performed to identify patientssuspected of having an anastomotic leak between January2002 and December 2008. Patients were identified from aradiology reports database and from clinical coding (ICD-10K918, K914). A leak was defined as a breakdown of acolonic anastomosis associated with an intra-abdominalcollection identified either by contrast radiographs or by thesurgeon at the time of a subsequent operation (modifiedfrom Damrauer [5]). Leaks were further divided intogeneralised leaks or contained leaks based on operativefindings or radiological appearance.

A standardised pro forma was used to collect data onpatient demographics, diagnosis and management, surgicalinterventions, morbidity and mortality and stoma rates.Patients with incidental leaks identified radiologically, andthose requiring no intervention were included. Primaryoutcome measures were: 1) mode of diagnosis; 2) day ofleak; 3) management of leak; 4) mortality; 5) stoma reversaland 6) demand on hospital resources.

Case notes were then audited, and we identified 30patients (male=16, female=14) with an anastomotic leak.Median age was 60 years (range 25–84). Mean ASA was 2(range 1–3) and mean BMI 27.1% (22/30 patients withBMI >25). Eight out of 30 patients (26.6%) were smokers.The length of follow up was between 18 months and7 years. The primary indication for surgery and the surgicalprocedure performed are listed in Table 1. Seven patientswho leaked following anterior resection had pre-operativeradiotherapy. Twenty out of the 30 patients were operatedby consultant-grade surgeons. Only one procedure wasperformed as an emergency. The anastomosis was stapledin 29 out of 30 cases.

For analysis, patients were grouped into two categories;those who leaked from a rectal anastomosis and the secondgroup who leaked from a colonic anastomosis above theperitoneal reflection. Leak rate was not assessed.

Results

Mode of diagnosis of anastomotic leak

Five patients were diagnosed on clinical grounds alone andproceeded directly to theatre for re-exploration. Fifteenpatients had a CT scan and five had a contrast enema priorto diagnosis being established.

A further five cases of leak were identified incidentallyin defunctioned patients following anterior resection at thetime of contrast study prior to planning elective stomareversal.

Time to diagnosis of anastomotic leak

The median time to presentation of clinically suspected leakswas 12 days (range 3–56 days). All clinically suspected leakswere identified during the primary admission. Of these,generalised leaks presented earliest (median day 8), followedby faecal fistula (median day 11) and localised leakspresenting latest (median day 18).

Clinically suspected leaks were detected from a rectalanastomosis at 14.7 days (range 4–32) and from a colonicanastomosis at 10.9 days (range 3–56).

Management of the anastomotic leak

Of the 14 patients who leaked from a rectal anastomosis(Fig. 1), ten had received a peri-operative defunctioningstoma. Only one generalised leak occurred in a patientwhose anastomosis was not protected by a covering stoma,the remainder being contained leaks (n=12) or faecal fistula(n=1).The proportion of patients requiring surgery in thedefunctioned group was lower (3/10 patients) compared tothe non-defunctioned group (3/4 patients). Of note, 70% (n=

Procedure No. of patients (n=30) Indication

Anterior resection Cancer

High 4

Low 9

Right hemicolectomy 10 2 IBD, 1 benign obstruction, 7 cancer

Left hemicolectomy 2 Cancer

Reversal of stoma 4 Elective reversal

Sigmoid colectomy 1 Cancer

Table 1 Type of procedure andindication for patients who suf-fered anastomotic leak

314 Int J Colorectal Dis (2011) 26:313–320

7) of patients in the defunctioned group needed no furtherintervention. Five patients whose leaks were detectedincidentally had all undergone anterior resection withdefunctioning stoma; none of them required any furtherintervention other than a delayed reversal of their stoma atsuch time as the leak was deemed to have healed by follow-up contrast study.

Of the 16 patients with a colonic anastomosis (Fig. 2),eight patients had a generalised leak with peritonitis andrequired surgery with six out of these, needing theanastomosis taking down with formation of end stoma. Intwo cases, following generalised leaks from a right hemi-colectomy, an initial attempt at repair of the anastomosiswas made. Both subsequently re-leaked and these patientsdied. Six patients with a contained leak were managed non-operatively with either antibiotics or percutaneous drain;

two of these were too unwell to re-operate, one had acontained leak after a right hemicolectomy and received apercutaneous drain and total parenteral nutrition (TPN), andtwo patients had small leaks that responded to antibioticsand a period of TPN. In two patients, the leak manifested asa faecal fistula, which, in one patient, did not resolve withconservative measures and resulted in the anastomosisbeing taken down.

Mortality after anastomotic leak

The overall mortality in our series was 26.7% (8/30 patients),with a 30-day mortality of 16.7% (5/30 patients). Subgroupanalysis reveals a mortality of 7.1% (1/14 patients) followinga leak from a rectal anastomosis and 43.8% (7/16 patients)from a colonic anastomosis (Figs. 1 and 2,). The only

Fig. 1 Treatment of 14 patientswith anastomotic leak followingrectal anastomosis

Fig. 2 Treatment of 16 patientswith anastomotic leak followinga colonic anastomosis

315Int J Colorectal Dis (2011) 26:313–320

mortality in the rectal anastomosis group was in a patientwho did not have a covering stoma (see Fig. 1). No patientwith an incidental leak died. In the colonic anastomosisgroup, 50% (4/8 patients) of patients with a generalised leakrequiring surgery died. Failure to take down the anastomosis,following a generalised leak, had an unfavourable prognosiswith a mortality of 100% (2/2 patients). Also in this group,three deaths occurred in patients where attempts were madeto manage a contained leak conservatively.

Stoma outcomes

The overall permanent stoma rate, following an anastomot-ic leak in our series, was 23.3% (7/30 patients), excludingpatients who died. Stoma outcomes are shown in Fig. 3. Inthose patients who required the anastomosis to be takendown following a leak, the permanent stoma rate was62.5% (5/8 patients) in those who survived. In the rectalanastomosis group, the permanent stoma rate was 27% (3/11 patients), compared with 57.1% (4/7 patients) in thecolonic anastomosis group.

Reasons for non-reversal of stoma were patient choice (n=3), high anaesthetic risk (n=2), presence of metastaticdisease (n=1), perineal fistula (n=1).

Demand on hospital resources

These 30 patients spent a total of 1,205 days in hospital,with an average stay of 40 days. Thirteen patients requireda critical care bed for an average of 23 days. The demandson enteral and parenteral nutrition, and use of radiologicalimaging are also demonstrated in Table 2.

Further analysis reveals that patients requiring intensivecare treatment had a 53.8% mortality (7/13 patients),compared with a 5.8% mortality (1/17 patients) who weresuitable for ward management.

Discussion

This retrospective audit of anastomotic leaks aimed toreview the management and outcomes of anastomotic leaks

Fig. 3 Stoma outcomes follow-ing anastomotic leak

316 Int J Colorectal Dis (2011) 26:313–320

in patients undergoing colorectal surgery in our department.Our results show that the patients experiencing leaks werenot exceptional, with a broad age range, mean ASA of two,different primary surgeries and only one of the leakscoming from an emergency procedure. Of note, 73.3% ofpatients fell into the obese category as determined by aBMI>25. This may be significant as obesity is a recognisedrisk factor for anastomotic leaks. We were pleased toobserve the number of leaks occurring in patients withinflammatory bowel disease was low (2/30 patients),despite our unit's particular interest in the surgical manage-ment of these patients. The purpose of the study was toreport what happens to patients who experience ananastomotic leak after colorectal surgery, and assessmentof a leak rate was not our intention.

Intense focus has been placed on rectal cancer surgery,as well as high quality training over the last two decades,and this has led to a standardised operative approach andmethod of anastomosis. This may explain why we observedfewer leaks from a rectal anastomosis (n=14) in this studythan from a colonic anastomosis (n=16). The need for morefocus on the quality of colonic surgery is recognised, andsteps towards this are now being made from an oncologicalviewpoint [12].

The majority (80%) of our patients, suspected of havinga leak based on clinical assessment, underwent furtherimaging to confirm a diagnosis. Choice of investigationwas based on the surgeon's own preference, the conditionof the patient, the location of the anastomosis andavailability of radiological support. Twenty percent of thosepresenting clinically were diagnosed by examination alone.CT has been demonstrated to be superior to contrast studiesin diagnosis of anastomotic leak, whereas contrast enema isrecommended to assess healing of a colorectal anastomosisprior to stoma reversal [13, 14].

The average time to diagnosis of a clinically apparentleak was 12 days. This is somewhat later than thetraditionally taught 5–7 days, but supports the findings ofa recent study by Hymen et al. who reported a mean of12.7 days to diagnosis using a large prospective database

[15]. In this era of enhanced recovery, whereby patientsmay be potentially discharged as early as post-operativeday 3 [16, 17], surgeons need to be aware that patients riskbeing considered for discharge before an anastomotic leakmay become clinically apparent. Anecdotal observationssuggest that on occasions, the early features of ananastomotic leak may be diagnosed as ‘prolonged ileus’or other sources of sepsis, such as occult chest or urinarytract infection. One series showed that 59% of leaks wereinitially treated as a cardiac complication [18].This can leadto a delay of several days until a diagnosis is confirmed byCT or becomes more clinically apparent as evidenced byperitonitis and sepsis, by which time mortality is likely tobe greater [19].

If surgery was required, those with a leaking colonicanastomosis were more likely to have the anastomosistaken down and a diverting stoma formed (75%) thanafter leak from the rectum (30%). More generalised leaksoccurred from a colonic anastomosis than from a rectalanastomosis, and always resulted in surgery. In thisseries, nearly all mortality was in the group who leakedfrom a colonic anastomosis rather than from a rectalanastomosis (89% vs. 11%, respectively). An anastomo-sis performed after rectal resection was likely to bedefunctioned, and this is reflected in the observednumber of asymptomatic or incidental leaks in this group(50%), as well as reduced need for surgical intervention(0%) and lower mortality (0%). Our results support thefindings of previous studies that have shown defunction-ing stomas reduce mortality, permanent stoma rates andthe need for re-operation [20–22], even if actual leak ratemay not have been affected [23].

Patients with a colonic anastomotic leak managed non-operatively by percutaneous drain or by antibiotics alonehad a high mortality of 42.9%. This shows that conserva-tive measures alone may not always be successful andfailure of sepsis to resolve rapidly in these patients shouldlead to the decision to operate without further delay. Forcontained leaks, in patients not demonstrating features offrank peritonitis, minimally invasive techniques such aspercutaneous CT-guided drainage may seem like anattractive alternative to surgery for both the surgeon andthe patient as intestinal continuity can be maintained.Damrauer et al. showed that patients with contained leaksdiffered very little from those with free leaks in rates ofoperative intervention, morbidity or mortality; thus support-ing the role of early surgery even in contained leaks [5]. Inour study, mortality was lower after a generalised leak if theanastomosis could be taken down, and end stoma formed(33% mortality) compared to an attempt at repair (100%mortality).

The mortality in this series of 27% demonstrates that inthe twenty-first century, despite the advances over the last

Table 2 Use of hospital resources in patients who suffered a colonicanastomosis leak

Number Mean Range Total

Days in critical care 13 23 1–79 304

TPN (days) 15 18 2–55 276

Enteral feed (days) 6 28 6–64 168

Total no. of further surgeries 17 2 1–5 29

CT scans 21 2 1–4 37

Contrast studies 13 2 1–4 26

Total stay (days) 30 40 6–144 1205

317Int J Colorectal Dis (2011) 26:313–320

decade in critical care and radiological support, ananastomotic leak remains a devastating complication ofcolorectal surgery. It is interesting that this figure is littledifferent from the mortality rate in the Large Bowel CancerProject back in 1980 (22%) [6], despite a reduction inreported leak rates over this time from 51% in rectalanastomosis reported by Goligher in 1970[24], to 13% seenin the Large Bowel Cancer Project [6], to the frequentlyreported rates of less than 5% seen today [1, 7, 15, 25, 26].Advances in surgical training, techniques and equipmenthave allowed today's colorectal surgeons to perform moredemanding surgeries, with lower anastomoses in higher riskpatients than our colleagues were able to in the past.However, what modern medicine cannot overcome are thegenetic polymorphisms that determine the innate immuneresponse to sepsis [27–29], that in some patients may leadto an early post-leak mortality that is difficult to reduce.Delay in diagnosis is likely to be important as a prolongedtime with faecal contamination of the peritoneal cavity mayhave a significant effect on the cascade of events that canlead to multi-organ failure and death. In this series, morethan half the patients who required intensive care died.

One observation that deserves further comment is themortality after a leak from a right hemicolectomy of 60% inthis series. This procedure is often regarded as a trainingprocedure delegated to registrars, but the consequences offailure of the anastomosis are significant. An ileocolicanastomosis may be generally considered as quite robust inits healing properties due to good blood supply and relativelack of tension on the anastomosis. This may give a false senseof security in the recovery period and lead to delay indiagnosis. Indeed in this series, the average time to diagnosisof leak after a right hemicolectomy was 12.7 days, above theaverage for the non-rectal anastomosis group (10.9 days). TheNational Bowel Cancer Audit [30], NCEPOD [31], and others[32], have also shown that mortality is relatively high afterright hemicolectomy compared to other colonic resections,suggesting that a right hemicolectomy is more dangerous thancommonly perceived, but this may be accounted for by alarger proportion of emergency surgery in this group.

Patients who underwent faecal diversion for intra-abdominal sepsis, following anastomotic leak, had a high rateof permanent stoma (62.5%). Reversal rate following electivediversion for anterior resection was better with 70% reversed,probably due to the fact that most of these were asymptomaticpatients with leaks detected incidentally. Other groups havereported similar permanent stoma rates after anastomotic leak,and this shows that despite modern technical advances,definitive surgical management of a leaking anastomosisoften requires faecal diversion to preserve life, with attemptsat repair of an anastomosis often proving unwise with a highmortality. In this series, two unsuccessful attempts were madeto repair a leaking anastomosis and the patients died. Evidence

also suggests that patients who later undergo restoration ofintestinal continuity are at a higher risk of suffering a furtherleak [11]. There is some debate about the best operativemanagement of a leak. One strategy in minor leaks is asuture repair with distal colonic irrigation, peritoneal lavageand a proximal loop stoma, although this technique was notemployed in this study. This technique may reduce thechance of permanent stoma although there is little evidenceto support this.

The demands on critical care, radiology, nutritional supportand the length of stay was examined and, although somewhatexpected, the results are quite remarkable, especially when itis observed that 30 patients occupied a hospital bed for morethan 3 years in total. This contrasts with the impressive resultsobtainable by enhanced recovery protocols. The managementof anastomotic leaks places an enormous burden on hospitalresources with additional time devoted to these patients bynursing staff and surgeons and other health care professionals[33]. Additional costs are likely to be considerable but nostudy has tried to quantify this.

Aweakness of our study is the retrospective nature of theanalysis, which may have led us to underestimating thenumber of leaks. Furthermore, we are only offering aninsight into our local management of anastomotic leaks andcannot, therefore, advise on best management. A source ofbias may be if any patients subsequently represented with aleak to another unit. However, coding data and theradiological reports database are collected prospectively,and all cases of suspected leak were subjected to a thoroughcase note and electronic record review. We also employedstrict diagnostic criteria when defining a leak as a breakdownof a colonic anastomosis associated with an intra-abdominalcollection identified either by contrast radiographs or by thesurgeon at the time of a subsequent operation. This accuratedefinition distinguishes a leak from an abscess adjacent to ananastomosis due to bacterial contamination not in continuitywith the bowel lumen. Different definitions have caused awide range of reported leak rates and make comparisondifficult between studies [34].

Although certain background risk factors for anastomoticleak do exist, such as nutritional status, alcohol intake,smoking, ASA score, BMI, rectal location of disease, surgicaltechnique remains an important consideration, as suggested bythe wide variation in leak rates between surgeons. Tension atthe anastomosis, poor blood supply or inadequate sutureplacement can all contribute to failure of an anastomosis. Airleak testing after colorectal resection can help reduce thefrequency of clinical leaks and should routinely be employed[26]. Novel adjuncts to anastomotic technique, such ascolorectal seals [35], polyester stents [36] and anastomoticbuttressing with absorbable materials [37], all show promisein reducing anastomotic leaks, but their routine use is someway off, given the lack of extensive comparative data [38, 39].

318 Int J Colorectal Dis (2011) 26:313–320

Prevention is always better than cure, but if a leak doesoccur, rapid diagnosis and definitive management withtaking down of the anastomosis and diverting stoma may bea strategy that confers a survival advantage.

Conflict of interest None to declare.

References

1. Buchs NC, Gervaz P, Secic M, Bucher P, Mugnier-Konrad B,Morel P (2008) Incidence, consequences, and risk factors foranastomotic dehiscence after colorectal surgery: a prospectivemonocentric study. Int J Colorectal Dis 23(3):265–270.doi:10.1007/s00384-007-0399-3

2. Mella J, Biffin A, Radcliffe AG, Stamatakis JD, Steele RJ (1997)Population-based audit of colorectal cancer management in twoUK health regions. Colorectal Cancer Working Group, RoyalCollege of Surgeons of England Clinical Epidemiology and AuditUnit. Br J Surg 84(12):1731–1736

3. Smith JA, King PM, Lane RH, Thompson MR (2003) Evidenceof the effect of ‘specialization’ on the management, surgicaloutcome and survival from colorectal cancer in Wessex. Br J Surg90(5):583–592. doi:10.1002/bjs.4085

4. Sorensen LT, Jorgensen T, Kirkeby LT, Skovdal J, Vennits B,Wille-Jorgensen P (1999) Smoking and alcohol abuse are majorrisk factors for anastomotic leakage in colorectal surgery. Br JSurg 86(7):927–931. doi:10.1046/j.1365-2168.1999.01165.x

5. Damrauer SM, Bordeianou L, Berger D (2009) Containedanastomotic leaks after colorectal surgery: are we too slow toact? Arch Surg 144(4):333–338. doi:10.1001/archsurg.2008.589

6. Fielding LP, Stewart-Brown S, Blesovsky L, Kearney G (1980)Anastomotic integrity after operations for large-bowel cancer: amulticentre study. Br Med J 281(6237):411–414

7. Branagan G, Finnis D (2005) Prognosis after anastomotic leakagein colorectal surgery. Dis Colon Rectum 48(5):1021–1026.doi:10.1007/s10350-004-0869-4

8. Law WL, Choi HK, Lee YM, Ho JW, Seto CL (2007)Anastomotic leakage is associated with poor long-term outcomein patients after curative colorectal resection for malignancy. JGastrointest Surg 11(1):8–15. doi:10.1007/s11605-006-0049-z

9. McArdle CS, McMillan DC, Hole DJ (2005) Impact of anasto-motic leakage on long-term survival of patients undergoingcurative resection for colorectal cancer. Br J Surg 92(9):1150–1154. doi:10.1002/bjs.5054

10. Nesbakken A, Nygaard K, Lunde OC (2001) Outcome and latefunctional results after anastomotic leakage following mesorectalexcision for rectal cancer. Br J Surg 88(3):400–404. doi:10.1046/j.1365-2168.2001.01719.x

11. Francone TD, Saleem A, Read TA, Roberts PL, Marcello PW,Schoetz DJ, Ricciardi R (2010) Ultimate fate of the leakingintestinal anastomosis: does leak mean permanent stoma? JGastrointest Surg 14(6):987–992. doi:10.1007/s11605-010-1190-2

12. West NP, Hohenberger W, Weber K, Perrakis A, Finan PJ, QuirkeP (2010) Complete mesocolic excision with central vascularligation produces an oncologically superior specimen comparedwith standard surgery for carcinoma of the colon. J Clin Oncol 28(2):272–278. doi:10.1200/JCO.2009.24.1448

13. Alves A, Panis Y, Pocard M, Regimbeau JM, Valleur P (1999)Management of anastomotic leakage after nondiverted large bowelresection—a multicentre study. J Am Coll Surg 189(6):554–559.doi:S1072-7515(99)00207-0

14. Lim M, Akhtar S, Sasapu K, Harris K, Burke D, Sagar P,Finan P (2006) Clinical and subclinical leaks after lowcolorectal anastomosis: a clinical and radiologic study. DisColon Rectum 49(10):1611–1619. doi:10.1007/s10350-006-0663-6

15. Hyman N, Manchester TL, Osler T, Burns B, Cataldo PA (2007)Anastomotic leaks after intestinal anastomosis: it's later than youth ink . Ann Surg 245(2) :254–258 . do i :10 .1097/01 .sla.0000225083.27182.85

16. Anderson AD, McNaught CE, MacFie J, Tring I, Barker P,Mitchell CJ (2003) Randomized clinical trial of multimodaloptimization and standard perioperative surgical care. Br J Surg90(12):1497–1504. doi:10.1002/bjs.4371

17. Eskicioglu C, Forbes SS, Aarts MA, Okrainec A, McLeod RS(2009) Enhanced recovery after surgery (ERAS) programs forpatients having colorectal surgery: a meta-analysis of randomizedtrials. J Gastrointest Surg 13(12):2321–2329. doi:10.1007/s11605-009-0927-2

18. Sutton CD, Marshall LJ, Williams N, Berry DP, Thomas WM,Kelly MJ (2004) Colo-rectal anastomotic leakage often masquer-ades as a cardiac complication. Colorectal Dis 6(1):21–22.doi:574

19. Macarthur DC, Nixon SJ, Aitken RJ (1998) Avoidable deaths stilloccur after large bowel surgery. Scottish Audit of SurgicalMortality, Royal College of Surgeons of Edinburgh. Br J Surg85(1):80–83. doi:10.1046/j.1365-2168.1998.00554.x

20. Huser N, Michalski CW, Erkan M, Schuster T, Rosenberg R,Kleeff J, Friess H (2008) Systematic review and meta-analysisof the role of defunctioning stoma in low rectal cancersurgery. Ann Surg 248(1):52–60. doi:10.1097/SLA.0b013e318176bf65

21. Karanjia ND, Corder AP, Bearn P, Heald RJ (1994) Leakage fromstapled low anastomosis after total mesorectal excision forcarcinoma of the rectum. Br J Surg 81(8):1224–1226

22. Tan WS, Tang CL, Shi L, Eu KW (2009) Meta-analysis ofdefunctioning stomas in low anterior resection for rectal cancer.Br J Surg 96(5):462–472. doi:10.1002/bjs.6594

23. Marusch F, Koch A, Schmidt U, Geibetaler S, Dralle H, SaegerHD, Wolff S, Nestler G, Pross M, Gastinger I, Lippert H (2002)Value of a protective stoma in low anterior resections for rectalcancer. Dis Colon Rectum 45(9):1164–1171. doi:10.1097/01.DCR.0000027040.59190.12

24. Goligher JC, Graham NG, De Dombal FT (1970) Anastomoticdehiscence after anterior resection of rectum and sigmoid. Br JSurg 57(2):109–118

25. Choi HK, Law WL, Ho JW (2006) Leakage after resection andintraperitoneal anastomosis for colorectal malignancy: analysis ofrisk factors. Dis Colon Rectum 49(11):1719–1725. doi:10.1007/s10350-006-0703-2

26. Ricciardi R, Roberts PL, Marcello PW, Hall JF, Read TE, SchoetzDJ (2009) Anastomotic leak testing after colorectal resection:what are the data? Arch Surg 144(5):407–411. doi:10.1001/archsurg.2009.43, discussion 411–402

27. Sorensen TI, Nielsen GG, Andersen PK, Teasdale TW (1988)Genetic and environmental influences on premature death in adultadoptees. N Engl J Med 318(12):727–732

28. Sutherland AM, Walley KR (2009) Bench-to-bedside review:association of genetic variation with sepsis. Crit Care 13(2):210.doi:10.1186/cc7702

29. Villar J, Maca-Meyer N, Perez-Mendez L, Flores C (2004) Bench-to-bedside review: understanding genetic predisposition to sepsis.Crit Care 8(3):180–189. doi:10.1186/cc2863

30. Smith JJ, Cornish J, Tekkis PP, Thompson MR (2007) TheNational Bowel Cancer Audit Project 2007: Quality improvementand open reporting. Association of Coloproctology of GreatBritain and Ireland, London

319Int J Colorectal Dis (2011) 26:313–320

31. Callum KG, Gray AJG, Hoile RW, Ingram GS, Martin IC,Sherry KM, Whimster F (2000) The 2000 report of thenational enquiry into perioperative deaths: then and now.NCEPOD, London

32. Boyd JB, Bradford B Jr, Watne AL (1980) Operative risk factorsof colon resection in the elderly. Ann Surg 192(6):743–746

33. Frye J, Bokey EL, Chapuis PH, Sinclair G, Dent OF (2009)Anastomotic leakage after resection of colorectal cancer generatesprodigious use of hospital resources. Colorectal Dis 11(9):917–920. doi:10.1111/j.1463-1318.2008.01728.x

34. Bruce J, Krukowski ZH, Al-Khairy G, Russell EM, Park KG(2001) Systematic review of the definition and measurement ofanastomotic leak after gastrointestinal surgery. Br J Surg 88(9):1157–1168. doi:10.1046/j.0007-1323.2001.01829.x

35. Ghitulescu GA, Morin N, Jetty P, Belliveau P (2003) Revisitingthe biofragmentable anastomotic ring: is it safe in colonic surgery?Can J Surg 46(2):92–98

36. Tsereteli Z, Sporn E, Geiger TM, Cleveland D, Frazier S,Rawlings A, Bachman SL, Miedema BW, Thaler K (2008)Placement of a covered polyester stent prevents complicationsfrom a colorectal anastomotic leak and supports healing: random-ized controlled trial in a large animal model. Surgery 144(5):786–792. doi:10.1016/j.surg.2008.05.018

37. Yo LS, Consten EC, Quarles van Ufford HM, Gooszen HG,Gagner M (2006) Buttressing of the staple line in gastrointestinalanastomoses: overview of new technology designed to reduceperioperative complications. Dig Surg 23(5–6):283–291.doi:10.1159/000096648

38. Ho YH, Ashour MA (2010) Techniques for colorectal anastomosis.World J Gastroenterol 16(13):1610–1621

39. Madbouly KM, Hussein A, Omar W, Farid M (2010) Regeneratedoxidized cellulose reinforcement of low rectal anastomosis: do westill need diversion? Dis Colon Rectum 53(6):889–895.doi:10.1007/DCR.0b013e3181d3206e

320 Int J Colorectal Dis (2011) 26:313–320