Stapled Versus Handsewn Methods For Ileocolic Anastomoses (Review)

Stapled versus handsewn methods for ileocolic anastomoses (Review)
Choy PYG, Bissett IP, Docherty JG, Parry BR, Merrie A, Fitzgerald A
This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2011, Issue 9 http://www.thecochranelibrary.com
Stapled versus handsewn methods for ileocolic anastomoses (Review) Copyright 2011 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
TABLE OF CONTENTS HEADER . . . . . . . . . . . . . . . . . . . . . . . . . . . ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . PLAIN LANGUAGE SUMMARY . . . . . . . . . . . . . . . . . . BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 1. . . . . . . . . . . . . . . . . . . . . . . . . . Figure 2. . . . . . . . . . . . . . . . . . . . . . . . . . Figure 3. . . . . . . . . . . . . . . . . . . . . . . . . . Figure 4. . . . . . . . . . . . . . . . . . . . . . . . . . Figure 5. . . . . . . . . . . . . . . . . . . . . . . . . . DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . AUTHORS CONCLUSIONS . . . . . . . . . . . . . . . . . . . ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . CHARACTERISTICS OF STUDIES . . . . . . . . . . . . . . . . . DATA AND ANALYSES . . . . . . . . . . . . . . . . . . . . . . Analysis 1.1. Comparison 1 All studies, Outcome 1 Overall anastomotic leak. . . Analysis 1.2. Comparison 1 All studies, Outcome 2 Clinical anastomotic leak. . . Analysis 1.3. Comparison 1 All studies, Outcome 3 Radiological anastomotic leak. Analysis 1.4. Comparison 1 All studies, Outcome 4 Anastomotic stricture. . . . Analysis 1.5. Comparison 1 All studies, Outcome 5 Anastomotic haemorrhage. . Analysis 1.6. Comparison 1 All studies, Outcome 6 Anastomotic time (min). . . Analysis 1.7. Comparison 1 All studies, Outcome 7 Re-operation. . . . . . . Analysis 1.8. Comparison 1 All studies, Outcome 8 Operative mortality. . . . . Analysis 1.9. Comparison 1 All studies, Outcome 9 Intra-abdominal abscess. . . Analysis 1.10. Comparison 1 All studies, Outcome 10 Wound infection. . . . . Analysis 1.11. Comparison 1 All studies, Outcome 11 Length of hospital stay (day). Analysis 2.1. Comparison 2 Cancer, Outcome 1 Overall anastomotic leak. . . . Analysis 2.2. Comparison 2 Cancer, Outcome 2 Clinical anastomotic leak. . . . Analysis 2.3. Comparison 2 Cancer, Outcome 3 Radiological anastomotic leak. . Analysis 2.4. Comparison 2 Cancer, Outcome 4 Anastomotic time. . . . . . Analysis 2.5. Comparison 2 Cancer, Outcome 5 Operative mortality. . . . . . Analysis 2.6. Comparison 2 Cancer, Outcome 6 Intra-abdominal abscess. . . . Analysis 2.7. Comparison 2 Cancer, Outcome 7 Wound infection. . . . . . . Analysis 2.8. Comparison 2 Cancer, Outcome 8 Length of stay. . . . . . . . Analysis 3.1. Comparison 3 Non-Cancer, Outcome 1 Overall anastomotic leak. . Analysis 3.2. Comparison 3 Non-Cancer, Outcome 2 Clinical anastomotic leak. . Analysis 3.3. Comparison 3 Non-Cancer, Outcome 3 Anastomotic stricture. . . Analysis 3.4. Comparison 3 Non-Cancer, Outcome 4 Anastomotic haemorrhage. . Analysis 3.5. Comparison 3 Non-Cancer, Outcome 5 Anastomotic time. . . . Analysis 3.6. Comparison 3 Non-Cancer, Outcome 6 Re-operation. . . . . . Analysis 3.7. Comparison 3 Non-Cancer, Outcome 7 Operative mortality. . . . Analysis 3.8. Comparison 3 Non-Cancer, Outcome 8 Intra-abdominal abscess. . Analysis 3.9. Comparison 3 Non-Cancer, Outcome 9 Wound infection. . . . . Analysis 3.10. Comparison 3 Non-Cancer, Outcome 10 Length of stay. . . . . WHATS NEW . . . . . . . . . . . . . . . . . . . . . . . . . HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . CONTRIBUTIONS OF AUTHORS . . . . . . . . . . . . . . . . . DECLARATIONS OF INTEREST . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1 1 2 2 3 3 5 6 6 7 7 8 8 9 10 10 12 25 26 27 28 28 29 29 30 31 32 33 34 35 36 36 37 38 38 39 40 40 41 42 42 43 43 44 44 45 46 46 46 47 47
i
SOURCES OF SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . INDEX TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
47 47
ii
[Intervention Review]
Stapled versus handsewn methods for ileocolic anastomoses

Pui Yee Grace Choy2 , Ian P Bissett1 , James G Docherty3 , Bryan R Parry4 , Arend Merrie5 , Anita Fitzgerald6 of Surgery, Auckland Hospital, Auckland, New Zealand. 2 Surgery, University of Auckland, Auckland, New Zealand. 3 Surgery, Raigmore Hospital, Inverness, UK. 4 Faculty of Medicine and Health Science, University of Auckland, Auckland, Australia. 5 Department of General Surgery, Auckland City Hospital, Auckland, New Zealand. 6 New Zealand Guidelines Group, Wellington, New Zealand Contact address: Ian P Bissett, Dept. of Surgery, Auckland Hospital, Parks Road, Auckland, 1, New Zealand. i.bissett@auckland.ac.nz. IanB@adhb.govt.nz. Editorial group: Cochrane Colorectal Cancer Group. Publication status and date: New search for studies and content updated (no change to conclusions), published in Issue 9, 2011. Review content assessed as up-to-date: 23 April 2011. Citation: Choy PYG, Bissett IP, Docherty JG, Parry BR, Merrie A, Fitzgerald A. Stapled versus handsewn methods for ileocolic anastomoses. Cochrane Database of Systematic Reviews 2011, Issue 9. Art. No.: CD004320. DOI: 10.1002/14651858.CD004320.pub3. Copyright 2011 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
1 Dept.
ABSTRACT Background Ileocolic anastomoses are commonly performed for right-sided colon cancer and Crohns disease. The anastomosis may be constructed using a linear cutter stapler or by suturing. Individual trials comparing stapled versus handsewn ileocolic anastomoses have found little difference in the complication rate but they have lacked adequate power to detect potential small difference. This is an update of a Cochrane review rst published in 2007. Objectives To compare outcomes of ileocolic anastomoses performed using stapling and handsewn techniques. The hypothesis tested was that the stapling technique is associated with fewer complications. Search methods MEDLINE, EMBASE, Cochrane Colorectal Cancer Group specialised register SR-COLOCA, Cochrane Library were searched for randomised controlled trials comparing use of a linear cuter stapler with any type of suturing technique for ileocolic anastomoses in adults from 1970 to 2005 and were updated in December 2010. Abstracts presented to the following society meetings between 1970 and 2010 were handsearched: American Society of Colon and Rectal Surgeons, the Association of Coloproctology of Great Britain and Ireland, European Association of Coloproctology. Selection criteria Randomised controlled trials comparing use of linear cutter stapler (isoperistaltic side to side or functional end to end) with any type of suturing technique in adults. Data collection and analysis Eligible studies were selected and their methodological quality assessed. Relevant results were extracted and missing data sought from the authors. RevMan 5 was used to perform meta-analysis when there were sufcient data. Sub-group analyses for cancer inammatory bowel disease as indication for ileocolic anastomoses were performed.
Stapled versus handsewn methods for ileocolic anastomoses (Review) Copyright 2011 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 1
Main results After obtaining individual data from authors for studies that include other anastomoses, seven trials (including one unpublished) with 1125 ileocolic participants (441 stapled, 684 handsewn) were included. The ve largest trials had adequate allocation concealment. Stapled anastomosis was associated with signicantly fewer anastomotic leaks compared with handsewn (S=11/441, HS=42/684, OR 0.48 [0.24, 0.95] p=0.03). One study performed routine radiology to detect asymptomatic leaks. For the sub-group of 825 cancer patients in four studies, stapled anastomosis led to signicantly fewer anastomotic leaks (S=4/300, HS=35/525, OR 0.28 [0.10, 0.75] p=0.01). In subgroup analysis of non-cancer patients (3 studies, 264 patients) there were no differences for any reported outcomes. All other outcomes: stricture, anastomotic haemorrhage, anastomotic time, re-operation, mortality, intra-abdominal abscess, wound infection, length of stay, showed no signicant difference. Authors conclusions Stapled functional end to end ileocolic anastomosis is associated with fewer leaks than handsewn anastomosis.
PLAIN LANGUAGE SUMMARY A comparison of two methods to join small and large bowel together in surgery Surgery for right-sided bowel cancer or Crohns disease commonly involve removing a segment of bowel and re-joining the small and large bowel together. The join, or anastomosis, can be made by stapling or sewing. This systematic review found seven randomised controlled trials with a total of 1125 participants (441 stapled, 684 handsewn) comparing these two methods. The leak rate from the bowel join for stapled anastomosis was 2.5%, signicantly lower than handsewn (6%). For the sub-group of 825 cancer patients in four studies, stapled join again has fewer leaks compared with handsewn, being 1.3% and 6.7% respectively. For the sub-group of 264 non-cancer patients in three studies, there were no differences for the reported outcomes. This sub-group included patients with Crohns disease. Overall, there was no signicant difference in the other outcomes of stricture, bleeding from the join, time to perform the join, re-operation, mortality, intra-abdominal abscess, wound infection and length of stay, although these were not consistently reported. The reason why a handsewn bowel join is more likely to leak is unclear. Possible explanations include less handling of the bowel, decreased spillage of bowel content during surgery, and uniform closure of all the staples using a stapler. This review did not compare different sewing materials or methods. The trials included in this review were performed from the early 1980s to 2009 involving six countries. The studies in Crohns disease were more recent but the combined number of patients was too small to summarise outcomes. More randomised controlled trials comparing the two surgical techniques in Crohns disease are needed.
BACKGROUND
Ileocolic anastomoses are commonly performed for right-sided colon cancer and Crohns disease. Colorectal cancer is the commonest gastrointestinal malignancy. It is the second leading cause of cancer deaths in the developed world. Right sided tumours account for around 30 % of colorectal cancer. The most common presentation of Crohns disease is in the ileum and caecum. Surgical treatment of these conditions requires right hemicolectomy or ileocaecal resection and formation of an ileocolic anastomosis. Two common methods to construct an anastomosis are the use of a linear cutter stapler and suturing.
Stapling devices have been in use since the late 1970s. A linear cutter stapler places two double staggered rows of staples and divides the tissue between the two lines of staples at the same time. There is however no consensus as to the superiority of stapling over handsewn methods for ileocolic anastomoses. The areas of contention regarding outcome include leak rate, stricture, speed of anastomosis formation and reoperation due to disease recurrence. The major problems associated with anastomoses are anastomotic leakage, stricture and bleeding (Brennan 1982). Contrast radiography is used in many studies to determine the true rate of leakage as many leaks are not manifested clinically. Stapled
2
anastomoses are generally thought to have a lower rate of leakage, both clinically and radiologically (Moran 1996). Conversely, some reports argued that overall leak rate is similar with handsewn anastomosis (Mann 1996). A higher rate of stricture formation is found with circular stapled anastomosis (MacRae 1998). However another study demonstrated that wide-lumen stapled anastomosis may decrease the incidence of recurrent Crohns disease (Munoz-Juarez 2001). Proponents of stapling devices cite speed but no study had found statistically signicant differences in operating time (Brundage 1999). The operating time with suturing was dependent on the type of sutures, for example, continuous sutures are quicker to insert than interrupted sutures. Stapling devices require familiarity with the instruments and are associated with a higher rate of technical mishap in the operating room (Moran 1996). Stapled devices are more expensive than sutures and this cost is not compensated by reduced hospital stay (Scher 1982). Comparison of the different anastomotic techniques is difcult as ileocolic anastomoses have to date not been analysed separately from other types of anastomoses or the exact numbers in the groups were not given. Of the studies that have been done sample size has been small giving insufcient power to draw valid conclusions. Thus there is uncertainty regarding current evidence. A systematic review to resolve these differences by pooling together and analysing all the available data was published in 2007 and updated in December 2010. The subject of colorectal anastomosis with circular stapler was analysed in a different review and will not be included (Lustosa 2002).
were included. If the trial included other types of gastrointestinal anastomosis, data for ileocolic anastomosis were extracted. If data were unavailable for this group, the authors were contacted to provide full details. Trials published after 1970 were included since stapling devices were not available in clinical use prior to this time. Types of participants Adults who received stapled and handsewn ileocolic anastomoses, elective or emergency. Subgroup analyses for cancer and inammatory bowel disease. Types of interventions Linear cutter stapler forming an ileocolic anastomosis either isoperistaltic side to side or functional end to end, the enterotomy remaining after withdrawal of the stapler may be closed with a stapling instrument or manual suturing; handsewn (any type of suturing material and technique) ileocolic anastomoses. Types of outcome measures Primary outcome: overall anastomotic leak - either clinical or radiological anastomotic leak Secondary outcomes: 1. Clinical anastomotic leak: associated with clinical signs and symptoms 2. Radiological anastomotic leak: anastomotic leak detected on the control postoperative enema, in a patient with no evidence of clinical anastomotic leak 3. Anastomotic stricture: narrowing of bowel lumen due to anastomotic healing 4. Anastomotic haemorrhage: postoperative rectal bleeding from anastomotic site 5. Anastomotic time: time required to perform the anastomosis 6. Re-operation: surgical intervention for complication or disease recurrence 7. Operative mortality: within 30 days 8. Intra-abdominal abscess 9. Wound infection 10. Length of hospital stay: time from operation to discharge from the hospital
OBJECTIVES
This review compares the use of linear cutter stapler and manual suturing in the formation of an ileocolic anastomosis. It aims to ascertain whether there is any difference in outcome between the two methods. It will also analyse the outcomes of the two techniques in subgroups of patients with cancer and inammatory bowel disease. The hypothesis to be tested is that stapling technique is associated with decreased level of complications.
METHODS Search methods for identication of studies Criteria for considering studies for this review
We planned to search for published and unpublished randomised controlled trials performed after 1970 with no restriction on language in the following electronic databases: MEDLINE, EMBASE, Cochrane Colorectal Cancer Group specialised register SR-COLOCA, Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews, Database of Abstracts
3
Types of studies All randomised controlled trials comparing the outcomes of linear cutter stapler and suture techniques for ileocolic anastomosis
of Reviews of Effectiveness for the years 1970 to 2010. This search strategy was based on MEDLINE and was adapted for each database search: 1. Collaborative Review Group search strategy for RCTs 2. anastom$.mp,hw. 3. ileocol$.mp,hw. or (ileum and colon$).mp,hw. 4. ((small adj2 intestin$) or (small adj2 bowel$)).mp,hw. 5. ((large adj2 intestin$) or (large adj2 bowel$)).mp,hw. 6. 4 or 5 7. (sutur$ or stapl$ or sew or sewn or handsew$ or stitch$).mp,hw. 8. anastomosis, surgical/ 9. intestine, large/ or colon/ 10. intestine, small/ or ileum/ 11. 9 or 10 12. exp sutures/ 13. exp suture techniques/ 14. surgical staplers/ 15. 2 or 8 16. 3 or 6 or 11 17. 7 or 12 or 13 or 14 18. 15 and 16 and 17 Search strategy for this review, performed 16.09.2005 by Karin Nielsen (CCCG) and updated by Margaret Paterson (New Zealand Guidelines Group) in December 2010. From the Cochrane Library: #1. (stapler or handsew* or stitch or (suture next techniques) or (surgical next staplers)) 947 #2. (anastom* or (anastomosis next sugical)) 1293 #3. (ileocol* or ileum or (large next intestin*) or (small next intestin*) or (large next bowel) or (small next bowel)) 1958 #4. (#1 and #2 and #3) 31 (23 hits in The Cochrane Central Register of Controlled Trials (CENTRAL)) (5 hits in The Cochrane Database of Systematic Reviews) (1 hit in Database of Abstracts of Reviews of Effects) EMBASE (Webspirs, Silver Platter version 2.0) (52 hits): (43 were saved) #6 #5 not (rabbit or dog or rat) 52 #5 (sutur* or stapl* or sew or sewn or handsew* or stitch or (suture techniques) or (surgical staplers)) and (anastom* or anastomosis, sugical) and ((ileocol*) or (large intestin*) or (small intestin*) or (large bowel) or (small bowel) or (ileum and colon)) and (random* or blind* or meta-analysis or placebo*) 72 #4 sutur* or stapl* or sew or sewn or handsew* or stitch or (suture techniques) or (surgical staplers) 33958 #3 anastom* or anastomosis, sugical 40321 #2 (ileocol*) or (large intestin*) or (small intestin*) or (large bowel) or (small bowel) or (ileum and colon) 108797 #1 random* or blind* or meta-analysis or placebo* 436277 MEDLINE (Webspirs, Silver Platter version 2.0) ( 58 hits): (34 saved)
#6 #5 not (rabbit or dog or rat or pig) 58 #5 (sutur* or stapl* or sew or sewn or handsew* or stitch or (suture techniques) or (surgical staplers)) and (anastom* or anastomosis, sugical) and ((ileocol*) or (large intestin*) or (small intestin*) or (large bowel) or (small bowel) or (ileum and colon)) and (random* or blind* or meta-analysis or placebo*) 65 #4 sutur* or stapl* or sew or sewn or handsew* or stitch or (suture techniques) or (surgical staplers) 62563 #3 anastom* or anastomosis, sugical 53899 #2 (ileocol*) or (large intestin*) or (small intestin*) or (large bowel) or (small bowel) or (ileum and colon) 93009 #1 random* or blind* or meta-analysis or placebo* 510192 Where possible principal authors were contacted for further information relating to the study and any other studies published and unpublished. All reference lists were checked for further studies. Abstracts presented to the following society meetings between 1970 and 2010 were handsearched: American Society of Colon and Rectal Surgeons, the Association of Coloproctology of Great Britain and Ireland, European Association of Coloproctology. Update searches run in December 2010 yielded 140 additional references.
Data collection and analysis

The titles and abstracts of articles found in the original search were screened by two independent reviewers (IB and PC); updated searches were carried out by two reviewers (AF and PC). Full text of eligible studies were obtained and each reviewer independently assessed whether the studies meet the inclusion and exclusion criteria. The excluded studies were recorded and the reasons for exclusion stated. Any difference of opinion was resolved by the third arbitrator (AM). The methodological quality of all studies eligible for the review were assessed independently. Each included trial was read for the following criteria which is presented in a table describing the included studies: concealed randomisation, technique of randomisation, time of randomisation (preoperatively, intraoperatively), number of randomised patients, number of patients not randomised and reasons for this, exclusion after randomisation, blinding of observer, blinding of outcome assessment, similarity between treatment and control group at entry, representativeness of patients, prospective data collection, dealing with drop outs, follow-up, standardisation of outcome assessment and whether an intention-to-treat analysis was performed. The presence of a learning-curve bias was assessed. The previous review utilised a scale assessing quality of randomisation (Jadad 1996). This scale was no longer used and replaced with Risk of bias tables. Two investigators independently extract the results of each trial on a standardised data sheet and to allow for cross-checking. Where possible, missing data were sought from the authors in the form of individual patient data. The software RevMan 5 provided by the Cochrane Collaboration was used for statistical analysis. Where
4
there were sufcient data, a summary statistic for each outcome was calculated. Where appropriate, formal meta-analysis and investigation of heterogeneity among trials were conducted. There were sub-group analyses for malignant disease and Crohns disease as the indication for ileocolic anastomoses using the same outcome measurements.
Risk of bias in included studies

The methodological details are shown in the Characteristics of included studies table. All seven studies were prospective randomised in design. Randomisation time was during the operation after the resection when either technique was deemed to be feasible and equally appropriate. In four trials, randomisation technique was sealed envelopes or concealment in sequential forms (Docherty 1995; Docherty 1991; Ikeuchi 2000; Kracht 1993) and one study used a phone-in randomisation method, which kept allocation concealed (McLeod 2009). These methods were considered adequate. The technique was not described for two studies, thus the allocation concealment was rated as unclear (Didolkar 1986; Izbicki 1998). The largest study (Kracht 1993) contributed 39% of patients . All studies had clearly dened inclusion & exclusion criteria. Assessment of randomisation success was done in all studies by showing that there was no signicant difference between most of the baseline patient characteristics. Three studies had excluded participants post randomisation for a variety of reasons including protocol violation and the palliative nature of surgery (Docherty 1995; Kracht 1993; McLeod 2009). Patients excluded post randomisation for palliative nature of surgery were included in the individual patient data meta-analysis (Docherty 1995). Follow-up duration ranged from 30 days post discharge (Kracht 1993) to a median of 87 months (Ikeuchi 2000). In one study (Docherty 1995), it was left to the discretion of surgeon, and in another (Didolkar 1986), standardised protocols for malignancy follow-up were used. Four studies mentioned lost to follow-up (Docherty 1995; Ikeuchi 2000; Izbicki 1998; McLeod 2009). Intention to treat analysis was reported in one study (Docherty 1995). Sample size was calculated in two studies (Docherty 1995; Kracht 1993). Two articles commented that the study situation reected daily surgical practice (Docherty 1995; Kracht 1993). Learning curve was taken into account in four studies in the selection of surgeons (Didolkar 1986; Docherty 1995; Kracht 1993; McLeod 2009). Only one study had blinded assessment of outcome by radiologist (Kracht 1993), while another was stated to be doubleblind but the blinding was not explained (Ikeuchi 2000).
RESULTS
Description of studies
See: Characteristics of included studies; Characteristics of excluded studies. Electronic search on MEDLINE, EMBASE and the Cochrane Library yielded 250 hits. Only 29 studies matched the inclusion criteria. The same data set was published in three articles on two occasions (Akyol 1991; Docherty 1995; West 1991; Yamamoto 1999a; Yamamoto 1999; Yamamoto 1999b), and in two articles on two occasions (Ikeuchi 2000; Kracht 1990; Kracht 1993; Kusunoki 1998). Eighteen studies were excluded mainly due to a lack of randomisation. One unpublished study (Docherty 1991) was obtained from the author of one of the included studies (Docherty 1995) through personal correspondence. It was unpublished due to problems with follow-up. This leaves six distinct trials for inclusion. Search of the SR-COLOCA by Trials Search Co-coordinator at CCCG and hand searching of the above mentioned meeting abstracts have not yielded other studies that were considered eligible for assessment. Searches were re-run in December 2010, resulting in the inclusion of one additional article (McLeod 2009) and two excluded studies. Only one series of patients looked at ileocolic anastomosis cases exclusively (Kracht 1993). For the remainder, ileocolic anastomosis was one of the different types of anastomoses studied. The total number of ileocolic participants were 1125 of which 441 (39%) were stapled and 684 (61%) were handsewn. This unequal distribution was due to the major study in this review having 4 different groups of handsewn anastomosis (Kracht 1993). More patients (100 patients) were included in the stapled group so that this group could be compared with the 4 sutured groups taken as a whole or individually (80 patients). Two studies examined anastomosis in Crohns disease patients only (Ikeuchi 2000; McLeod 2009) while another study excluded Crohns disease (Izbicki 1998). Three studies focused on patients with cancer (Didolkar 1986; Docherty 1991; Kracht 1993). The remaining study had no restriction on type of disease. Two studies excluded emergency operations (Izbicki 1998; Kracht 1993). Subgroup analysis excluded the study by Izbicki 1998 as individual patient data for cancer were not available. The new study included patients who had laparoscopic resections (McLeod 2009).
Effects of interventions
The main results from analysis of seven trials with a total of 1125 ileocolic anastomosis patients (Stapled=441, Handsewn=684) are shown here, see Data and analyses for further details. Individual patient data were used for Docherty 1991, Docherty 1995 & Ikeuchi 2000. (1) Overall anastomotic leak (7 studies, 1125 patients): stapled anastomosis (11/441, 2.5%) was associated with signicantly fewer leaks compared with handsewn (42/684, 6%) with odds ratio of 0.48 [95% condence interval 0.24, 0.95; p=0.03] (Figure
5
1). Figure 1. Forest plot of comparison: 1 All studies, outcome: 1.1 Overall anastomotic leak.
(2) Clinical anastomotic leak (7 studies, 1125 patients): stapled anastomosis (10/441, 2.3%) appeared to have fewer leaks compared with handsewn (29/684, 4.2%) with OR 0.55 [0.27, 1.15, p=0.11] however the difference between groups was not signicant (Figure 2). Figure 2. Forest plot of comparison: 1 All studies, outcome: 1.2 Clinical anastomotic leak.
(3) Radiological anastomotic leak (1 study, 440 patients) rate for stapled anastomosis was 0.9% (1/106) and handsewn was 3.9% (13/334), OR 0.24 [0.03, 1.82, p=0.17]. (4) Anastomotic stricture (2 studies, 65 patients): stapled anastomosis (1/26) & handsewn (6/39), occurrence too low to allow meta-analysis.
(5) Anastomotic haemorrhage (2 studies, 65 patients): no occurrences in the studies (6) Anastomotic time (1 study, 255 patients): though only one study, stapled anastomosis took an average of 8.72 min (standard deviation 5.12) to construct & handsewn 22.36 min (S.D. 11.54).
6
(7) Re-operation (2 studies, 199 patients): there were no differences in re-operation rates between stapled anastomosis (7/95, 7.3%) and handsewn (12/104, 11.5%) with and OR of 0.61 [0.25, 1.18, p=0.43]. (8) Operative mortality (6 studies, 1087 patients): minor variation in length of follow up but all included death within 30 days. For Docherty 1995, 2 out of 6 deaths in handsewn group had anastomotic leak. For Docherty 1991, all 3 deaths in handsewn group had anastomotic leak. None of the deaths in stapled group had anastomotic leak. In Kracht 1993, 4 of the 14 deaths had intra-abdominal sepsis, and were all in the handsewn group. Three studies reported no operative mortality in either group. (Ikeuchi 2000; Izbicki 1998; McLeod 2009)
(9) Intra-abdominal abscess (5 studies, 932 patients): There were no differences between stapled (4/356, 1.1%) and handsewn (19/ 576, 3.3%) groups with OR 0.41 [0.14, 1.25, p=0.12] (10) Wound infection (5 studies, 932 patients): There were no differences between stapled (33/356, 9.3%) and handsewn (53/ 576, 9.2%) groups, OR 1.05 [0.66, 1.70, p=0.83] (11) Length of hospital stay (3 studies, 424 patients): no difference between stapled & handsewn anastomoses, OR 0.19 [-1.50, 1.87, p=0.8] In subgroup analysis of cancer patients (4 studies, 825 patients): (a) Overall anastomotic leak rate was signicantly lower in stapled group (4/300, 1.3%) than handsewn group (35/525, 6.7%) with OR 0.28 [0.10, 0.75, p=0.01] (Figure 3).
Figure 3. Forest plot of comparison: 2 Cancer, outcome: 2.1 Overall anastomotic leak.
(b) Clinical anastomotic leak: Stapled anastomosis (3/300, 1%) was associated with signicantly fewer leaks compared with handsewn (22/525, 4.2%) with OR 0.30 [0.10, 0.95, p=0.04] (Figure 4). Figure 4. Forest plot of comparison: 2 Cancer, outcome: 2.2 Clinical anastomotic leak.
(c) Radiological anastomotic leak: the study that considered radiological leak only included cancer patients and is reported above. (d) Except for anastomotic time being shorter in stapled group (see above), operative mortality, intra-abdominal abscess, wound infection & length of stay showed no difference. In subgroup analysis of non-cancer patients (3 studies, 264 patients) there were no differences between groups for overall anastomotic leak (Figure 5), clinical anastomotic leak, re-operation, operative mortality, intra-abdominal abscess or wound infection. Other outcomes were not estimable as the number of participants was too small. Figure 5. Forest plot of comparison: 3 Non-Cancer, outcome: 3.1 Overall anastomotic leak.
DISCUSSION
This is the largest systematic review that specically investigates the relative results of handsewn and stapled ileocolic anastomoses. Ileocolic anastomosis needs to be analysed separately from other colonic anastomosis, as the involved bowel ends are different with regard to diameter, wall structure, location in the abdomen and bacterial characteristics (Kracht 1993). The stapled technique for ileocolic anastomosis differs from distal anastomoses and anastomotic leak rate, and is higher with left-sided anastomosis (Lipska 2006). Except for two studies (Kracht 1993; McLeod 2009), all studies included other colonic, small bowel, and/or upper gastrointestinal anastomoses. To enable direct comparison of outcome measures, individual patient data regarding ileocolic anastomoses were used. The aim of this review is to compare ileocolic anastomosis only, thus most other studies have the results affected by other types of anastomoses. A previous meta-analysis (MacRae 1998) which included other colon & rectal anastomoses showed no signicant difference between handsewn and stapled methods in terms of leak rate. It stated overall leak rate of 9% for both handsewn and stapled anastomoses, with odds ratio favoured slightly stapled anastomosis
(0.92, 0.69-1.12). Another meta-analysis (Simillis 2007) investigated the outcomes of different anastomotic types after resection for Crohns disease. It found reduced ileocolonic anastomotic leaks in other anastomotic congurations compared with end-to-end anastomosis (OR3.8, p=0.05). However the investigators noted potential for bias due to the retrospective nature of most of the included studies. In this systematic review, stapled ileocolic anastomosis was associated with fewer anastomotic leaks compared with handsewn overall and in the subgroup of cancer patients. None of the seven studies individually found a signicant difference in leak rate between the two groups. The overall leak rate was 2.5% for stapled and 6% for handsewn anastomoses. Although high it includes radiological leaks. The clinical leak rate was 2.3% for stapled and 4.2% for handsewn. The reported leak rate falls between the published rates of 0.5%-7.0% (Alves 2002; Brennan 1982; Chassin 1978; Isbister 2001; Leslie 2003; Lipska 2006; Scher 1982; Simillis 2007). It was imperative that each of the included studies used vigorous methodology. All studies were prospectively randomised after resection when either technique was deemed appropriate. Allocation concealment for Docherty 1991; Docherty 1995; Ikeuchi 2000; Kracht 1993 and McLeod 2009 were adequate as the randomisation technique was sealed envelopes or concealment in sequen8
tial forms or computer generated with a phone-in service. Kracht 1993 was the largest and it looked at four different methods of suturing. It was constructed such that more patients were included in the stapled group so that this group could be compared with the 4 sutured groups taken as a whole or individually. The leak rate of stapled anastomosis (2.8%) was less than handsewn varieties (8.4%) overall with an a posterior gamma risk of less than 5%. The other six trials had similar number of participants in the two groups. No signicant differences in baseline characteristics were found by authors of Docherty 1995; Ikeuchi 2000; Izbicki 1998 and McLeod 2009. The ve comparison groups in Kracht 1993 had comparable preoperative criteria except there were signicantly more women (n= 72) compared with men (n=34) in stapled group (p<0.02). There were however more women (n=261) then men (n=179) in the entire study. Allocation concealment was adequate using appropriately generated sequence of randomisation. The discrepancy may be due to individual group size rather than failure of randomisation as 2 of the sutured groups also had more women than men. Some studies (Lipska 2006; Walker 2004) have shown that women have a lower incidence of anastomotic leak compared with men but the authors did not comment on the contribution of gender distribution to the anastomotic leak rates. Kracht 1993 did speculate however that a signicantly lower rate of intra-operative septic spillage in the stapled group (p<0.02) was a theoretical advantage of the functional end to end stapling technique. Stapled anastomosis led to less tissue trauma and decreased chances of peritoneal contamination (Tewari 2005; Simillis 2007). Intra-operative septic conditions have been shown to be a risk factor for clinically signicant anastomotic leak (Alves 2002). In this systematic review, intra-abdominal abscess also appeared more common in handsewn anastomosis. There was a non signicant trend to more intra-abdominal abscess in the sutured group. Only one study performed routine post-operative gastrogran enema which picked up asymptomatic radiological leaks (Kracht 1993). For the other two large studies (Docherty 1991; Docherty 1995), radiology was used to conrm clinical suspicion of anastomotic leak. The asymptomatic leaks detected by radiology may however have clinical sequelae from a nding that even localised anastomotic leak is associated with diminished overall and cancerspecic survival (Ho 2010; Walker 2004). When individual patient data in the sub-group of 825 patients with cancer in 4 studies were analysed together, overall anastomotic leak rate for stapled anastomoses was 1.3% and again was signicantly lower than handsewn (6.7%). Moreover, stapled anastomosis (3/300, 1%) was associated with signicantly fewer clinical leaks compared with handsewn (22/525, 4.2%) with OR 0.30 [0.10, 0.95, p=0.04]. The main shortcoming of this systematic review is that it was not able to compare between specic types of handsewn anastomosis.
Many papers specify only the materials used and not the method of constructing the anastomosis. There are still controversies regarding which is the preferred handsewn technique. The study by Kracht 1993 attempted to address the inuence of suture technique on outcome. Each of these handsewn techniques had a trend to increased leak rate compared with stapled, although only the leak rate for end to end continuous suture reached statistical signicance alone. The reason why handsewn ileocolic anastomoses have an increased leak rate is unclear. Possible reason is that the increased rate of local spillage with handsewn is a contributory factor. The uniform closure of all the staples may also be important, as well as reduction in tissue manipulations and less inammation (Simillis 2007). Anastomotic time theoretically should be shorter with stapled. But only one study examined and conrmed this nding (Docherty 1995). Both Didolkar 1986 & Izbicki 1998 included anastomotic time with other anastomoses and both found stapled anastomoses slightly faster to perform. However it does not always translate into shorter operation time (Izbicki 1998; Scher 1982). However, McLeod 2009 found both the anastomotic time and duration of operation to be signicantly faster with stapled anastomosis. Only one study (Izbicki 1998) examined cost and they found stapled anastomoses were more expensive. Whether this expense can be offset by savings in theatre time & staff pay is dependent on systemic issues in the hospital. An in depth analysis of costs between the two methods has not been performed. This systematic review included studies performed from the early 1980s to 2009 from six countries, one being a multi-centre study spanning Canadian, American and British centres (McLeod 2009). The majority of excluded papers did not randomise between stapled and handsewn groups. Though statistically signicant results were found for the primary outcome of anastomotic leak, too few of the other outcomes were reported to allow meaningful meta-analysis. There were more recent investigations (McLeod 2009;Simillis 2007) in patients with Crohns disease, there was a relative paucity of randomised controlled trials in Crohns disease with too few anastomotic leaks to draw any conclusions. This systematic review provides evidence on whether to use a linear cutter stapler or suturing to construct an ileocolic anastomosis. It afrms that the stapling technique has the advantage of a lower anastomotic leak rate.
AUTHORS CONCLUSIONS Implications for practice

Stapled functional end to end ileocolic anastomosis is associated with fewer anastomotic leaks than handsewn anastomosis, and should be considered the standard against which all other techniques should be compared.
9
Implications for research

There is no clear evidence in Crohns disease. More randomised controlled trials on stapled versus handsewn ileocolic anastomosis in Crohns disease are needed.
ACKNOWLEDGEMENTS
Sue Foggin, Information Services Librarian, Philson Library, University of Auckland, Auckland, New Zealand, for assisting with search strategy.
REFERENCES
References to studies included in this review

Didolkar 1986 {published data only} Didolkar MS, Reed WP, Elias EG, Schnaper LA, Brown SD, Chaudhary SM. A prospective randomized study of sutured versus stapled bowel anastomoses in patients with cancer. Cancer 1986;57:456460. Docherty 1991 {unpublished data only} Docherty JG, Rankin E, Galloway DJ on behalf of the anastomotic study group. Anastomotic integrity and local recurrence after colorectal cancer surgery. (personal communication: JG Docherty 11/07/2005) 1991. Docherty 1995 {published and unpublished data} Docherty JG, McGregor JR, Akyol AM, Murray GD, Galloway DJ, West of Scotland and Highland Anastomosis Study Group. Comparison of manually constructed and stapled anastomoses in colorectal surgery. Annals of Surgery 1995;221(2):176184. Ikeuchi 2000 {published and unpublished data} Ikeuchi H, Kusunoki M, Yamamura T. Long-term results of stapled and hand-sewn anastomoses in patients with Crohns Disease. Digestive Surgery 2000;17:493496. Izbicki 1998 {published data only} Izbicki JR, Gawad KA, Ouirrenbach S, Hosch SB, Breid V, Knoefel WT, et al.Can stapled anastomosis in visceral surgery still be justied? A prospective controlled randomized study of the cost-effectiveness of hand-sewn and stapled anastomoses [Ist die Klammernaht in der Visceralchirurgie noch gerechtfertigt? Eine prospektiv kontrollierte, randomisierte Studie zur Kosteneffektivitat von Hand und Klammernaht]. Der Chirurg 1998;69: 72534. Kracht 1993 {published data only} Kracht M, Hay J-M, Fagniez P-L, Fingerhut A. Ileocolonic anastomosis after right hemicolectomy for carcinoma: stapled or hand-sewn?. International Journal of Colorectal Disease 1993;8:2933.
McLeod 2009 {published data only} McLeod RS, Wolff BG, Ross S, Parkes R, McKenzie M, Investigators of the CAST Trial. Recurrence of Crohns disease after ileocolic resection is not affected by anastomotic type: results of a multicenter, randomized, controlled trial. Diseases of the Colon & Rectum 2009;52(5):919927.
References to studies excluded from this review

Alessandroni 2009 {published data only} Alessandroni L, Bertolini R, Campanelli A, Capaldi M, Di Castro A, Mencacci R, et al.Role of anastomotic conguration in ileocolic resection for Crohns disease. Chirurgia Italiana 2009;61(1):2331. Anwar 2004 {published data only} Anwar S, Huges S, Eadie AJ, Scott NA. Anastomotic technique and survival after right hemicolectomy for colorectal cancer. Surgeon Journal of the Royal Colleges of Surgeons of Edinburgh & Ireland 2004;2(5):27780. Brennan 1982 {published data only} Brennan SS, Pickford IR, Evans M, Pollock AV. Staples or sutures for colonic anastomoses-a controlled clinical trial. British Journal of Surgery. 1982;69:722724. Brundage 1999 {published data only} Brundage SI, Jurkovich GJ, Grossman DC, Tong WC, Mack CD, Maier RV. Stapled versus sutured gastrointestinal anastomoses in the trauma patient. Journal of Trauma: Injury, Infection, & Critical Care. 1999;47(3):500-7; discussion 507-8. Bubrick 1991 {published data only} Bubrick MP, Corman ML, Cahill CJ, Hardy TG, Jr, Nance FC, Shatney CH. Prospective, randomized trial of the biofragmentable anastomosis ring. American Journal of Surgery 1991;161:136143. Cajozzo 1990 {published data only} Cajozzo M, Compagno G, DiTora P, Spallitta SI, Bazan P. Advantages and disadvantages of mechanical vs manual anastomosis in colorectal surgery. Acta Chir Scand 1990; 156:167169.
10
Chassin 1978 {published data only} Chassin JL, Rifkind KM, Sussman B, Kassel B, Fingaret A, Drager S. The stapled gastrointestinal tract anastomosis: incidence of postoperative complications compared with the sutured anastomosis. Annals of Surgery 1978;188:689696. Corman 1989 {published data only} Corman ML, Prager ED, Hardy TG, Jr, Bubrick MP, the Valtrac (BAR) Study Group. Comparison of the Valtrac biofragmentable anastomosis ring with conventional suture and stapled anastomosis in colon surgery: results of a prospective randomized clinical trial. Diseases of the Colon & Rectum 1989;32(3):183187. Demetriades 2002 {published data only} Demetriades D, Murrary JA, Chan LS, Ordonez C, Bowley D, Nagy KK, et al.Handsewn versus stapled anastomosis in penetrating colon injuries requiring resection: a multicenter study. Journal of Trauma: Injury, Infection, & Critical Care 2002;52(1):11721. Dyess 1990 {published data only} Dyess DL, Curreri PW, Ferrara JJ. A new technique for sutureless intestinal anastomosis. A prospective, randomized, clinical trial. The American Surgeon 1990;56: 7175. Hashemi 1998 {published data only} Hashemi M, Novell JR, Lewis AAM. Side-to-side stapled anastomosis may delay recurrence in Crohns Disease. Diseases of the Colon & Rectum 1998;41:12931296. Munoz-Juarez 2001 {published data only} Munoz-Juarez M, Yamamoto T, Wolff BG, Keighley MRB. Wide-lumen stapled anastomosis vs. conventional endto-end anastomosis in the treatment of Crohns disease. Diseases of the Colon & Rectum 2001;44(1):20-5; discussion 25-6. Reiling 1980 {published data only} Reiling RB, Reiling WA, Bernie WA, Huffer AB, Perkins NC, Elliott DW. Prospective controlled study of gastrointestinal stapled anastomoses. The American Journal of Surgery 1980;139:147-151, discussion 151-152. Resegotti 2005 {published data only} Resegotti A, Astegiano M, Farina E C, Ciccone G, Avagnina G, Giustetto A. Side-to-side stapled anastomosis strongly reduces anastomotic leak rates in Crohns disease surgery. Diseases of the Colon & Rectum 2005;48(3):4648. Scarpa 2004 {published data only} Scarpa M, Angriman I, Barollo M, Polese L, Ruffolo C, Bertin M, et al.Role of stapled and hand-sewn anastomoses in recurrence of Crohns disease. Hepato-Gastroenterology 2004;51(58):10537. Scarpa 2007 {published data only} Scarpa M, Ruffolo C, Bertin E, Polese L, Filosa T, Prando D, et al.Surgical predictors of recurrence of Crohns disease after ileocolonic resection.. International Journal of Colorectal Disease 2007;22:10611069.
Scher 1982 {published data only} Scher KS, Scott-Conner C, Jones CW, Leach M. A comparison of stapled and sutured anastomoses in colonic operations. Surgery, Gynecology & Obstetrics 1982;155: 489493. Tersigni 2003 {published data only} Tersigni R, Alessandroni L, Barreca M, Piovanello P, Prantera C. Does stapled functional end-to-end anastomosis affect recurrence of Crohns disease after ileocolonic resection?. Hepato-Gastroenterology 2003 SepOct;50(53): 14225. Wolmark 1986 {published data only} Wolmark N, Gordon PH, Fisher B, Weiand S, Lerner H, Lawrence W. A comparison of stapled and handsewn anastomoses in patients undergoing resection for Dukes B and C colorectal cancer: an analysis of disease-free survival and survival from the NSABP prospective clinical trials. Disease of the Colon & Rectum 1986;29:344350. Yamamoto 1999 {published data only} Yamamoto T, Bain IM, Mylonakis E, Allan RN, Keighley MRB. Stapled functional end-to-end anastomosis versus sutured end-to-end anastomosis after ileocolonic resection in Crohn disease. Scand J Gastroenterol 1999;34(7): 708713.
Additional references
Akyol 1991 Akyol AM, McGregor JR, Galloway DJ, Murray G, George WD. Recurrence of colorectal cancer after sutured and stapled large bowel anastomoses. British Journal of Surgery 1991;78(11):12971300. Alves 2002 Alves A, Panis Y, Trancart D, Regimbeau J-M, Pocard M, Valleur P. Factors associated with clinically signicant anastomotic leakage after large bowel resection: multivariate analysis of 707 patients. World Journal of Surgery 2002;26: 499502. Ho 2010 Ho Y-H, Ashour MAT. Techniques for colorectal anastomosis. World Journal of Gastroenterology 2010;16(13): 16101621. Isbister 2001 Isbister WH. Anastomotic leak in colorectal surgery: a single surgeons experience. ANZ Journal of Surgery 2001; 71:51620. Jadad 1996 Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJM, Gavaghan DJ, et al.Assessing the quality of reports of randomized clinical trials: is blinding necessary?. Controlled Clinical Trials 1996;17:112. Kracht 1990 Kracht M. Manual or mechanical right colic anastomoses [Anastomose colique droite manuelle ou mecanique?]. Chirurgie 1990;116:415418.
11
Kusunoki 1998 Kusunoki M, Ikeuchi H, Yanagi H, Shoji Y, Yamamura T. A comparison of stapled and hand-sewn anastomoses in Crohns. Digestive Surgery 1998;15:679682. Leslie 2003 Leslie A, Steele RJC. The interrupted serosubmucosal anastomosis - still the gold standard. Colorectal Disease 2003;5:3626. Lipska 2006 Lipska MA, Bissett IP, Parry, BR, Merrie AEH. Anastomotic leakage after lower gastrointestinal anastomosis: men are at a higher risk. ANZ Journal of Surgery 2006;76:579585. Lustosa 2002 Lustosa S, Matos D, Atallah AN, Castro AA. Stapled versus handsewn methods for colorectal anastomosis surgery. Cochrane Database of Systematic Reviews 2002, Issue 3. [DOI: 10.1002/14651858.CD003144] MacRae 1998 MacRae HM, McLeod RS. Handsewn vs.stapled anastomoses in colon and rectal surgery: a meta-analysis. Diseases of the Colon & Rectum 1998;41(2):180189. Mann 1996 Mann B, Kleinschmidt S, Stremmel W. Prospective study of hand-sutured anastomosis after colorectal resection. British Journal of Surgery. 1996;83:2931. Moran 1996 Moran BJ. Stapling instruments for intestinal anastomosis in colorectal surgery. British Journal of Surgery. 1996;83: 902909.
Simillis 2007 Simillis C, Purkayastha S, Yamamoto T, Strong SA, Darzi AW, Tekkis PP. A meta-analysis comparing conventional end-to-end anastomosis vs. other anastomotic congurations after resection in Crohns disease.. Diseases of the Colon & Rectum 2007;50(10):16741687. Tewari 2005 Tewari M, Shukla HS. Right colectomy with iso-peristaltic side-to-side stapled ileocolic anastomosis. Journal of Surgical Oncology 2005;89:99101. Walker 2004 Walker KG, Bell SW, Rickard MJFX, Mehanna D, Dent OF, Chapuis PH, et al.Anastomotic leakage is predictive of diminished survival after potentially curative resection for colorectal cancer. Annals of Surgery 2004;240(2):255259. West 1991 West of Scotland and Highland Anastomosis Study Group. Suturing or stapling in gastrointestinal surgery: a prospective randomized study. British Journal of Surgery 1991;78(3):337341. Yamamoto 1999a Yamamoto T, Allan RN, Keighley MR. Strategy for surgical management of ileocolonic anastomotic recurrence in Crohns disease. World Journal of Surgery 1999;23:1055-60, discussion 1960-1. Yamamoto 1999b Yamamoto T, Keighley MR. Stapled functional end-to-end anastomosis in Crohns disease. Surgery Today 1999;29: 67981. Indicates the major publication for the study
12
CHARACTERISTICS OF STUDIES Characteristics of included studies [ordered by study ID]

Didolkar 1986 Methods 1) Study location: Baltimore, Maryland 2) Study duration: not mentioned 3) Follow-up duration: not mentioned (long-term follow-up for their malignancies, often following standardised protocols, no reference given) 4) Randomisation technique: not mentioned 5) Randomisation time: in the OR after examination of the abdomen 6) Blinding: none 1) Source population: all patients in the Surgical Oncology Service requiring a large or small bowel anastomosis, under the care of 3 surgeons 2) Type of disease: any cancer 3) Type of anastomosis: small bowel-small bowel, colon-small bowel, colon-colon 4) Inclusion criteria: histologically proven cancer, one or more adverse factors for healing (advanced stage, carcinomatosis, prior radiation or chemotherapy, bowel obstruction, adhesions, infection, or steroid therapy) 5) Exclusion criteria: use of EEA stapler for low rectal anastomosis, gastric or oesophageal anastomoses 6) Number of participants: ileocolic n=38 (total in study, n=88) 7) Excluded post randomisation: not mentioned 8) Lost to follow-up: not mentioned 1) Group S: functional end-to-end with GIA and TA55 stapling instruments, ileocolic n=22 2) Group HS: a through-and-through layer of 3-0 chromic catgut and outer seromuscular layer of 3-0 silk suture, ileocolic n=16 1) Overall anastomotic leak: 0 2) Clinical anastomotic leak: 0 3) Intra-abdominal abscess: 0 4) Wound infection: 0 1) Randomisation success: preoperative laboratory values as a group comparable (except alkaline phosphatase higher in stapled, p=0.0543), other patient characteristics comparable except more female in stapled group, the incidence of patients with advanced cancer and abdominal carcinomatosis was higher in the stapled group (p=0.02), number of patients who underwent emergency operation was comparable 2) Representativity: not mentioned 3) Sample size: not calculated 4) Intention to treat: not done as no loss to follow-up is mentioned 5) Learning curve: all surgeons had three years of experience in stapling techniques 6) Others: stratication for prior radiation therapy, presence of bowel obstruction, infection, and the site of bowel anastomosis 7) individual data for ileocolic group not available as author (MS Didolkar) no longer keeps the les
Participants
Interventions
Outcomes
Notes
13
Didolkar 1986
(Continued)
Risk of bias Bias Authors judgement Support for judgement No blinding, but the review authors judge that the outcome and the outcome measurement are not likely to be inuenced by lack of blinding Reasons for missing outcome data unlikely to be related to true outcome Reported all pre-specied outcomes
Blinding (performance bias and detection Low risk bias) Subjective outcomes
Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias)
Low risk
Low risk
Docherty 1991 Methods 1) Study location: Scotland 2) Study duration: August 1991 to mid 1997. 3) Follow-up duration: planned to be 2 years 4) Randomisation technique: batch of sealed envelopes 5) Randomisation time: during operation when either technique would be equally feasible 6) Blinding: not described 1) Source population: all patients under care of participating consultants with diagnosis of colorectal cancer, elective or emergency 2) Type of disease: colorectal cancer 3) Type of anastomosis: stapled or handsewn colonic & rectal anastomoses 4) Inclusion criteria: providing either stapled or handsewn techniques are feasible 5) Exclusion criteria: if one technique is not feasible then no randomisation but remain in study 6) Number of participants: n=157 7) Excluded post randomisation: not mentioned 8) Lost to follow-up: not mentioned 1) Group S: Ethicon functional end-to-end staplers, n=70 2) Group HS: Ethicon sutures, n=87 1) Overall anastomotic leak: S=0, HS=4 2) Clinical anastomotic leak: S=0, HS=4 3) Operative mortality: S=2, HS=3 4) Length of stay: S=11.36 days (5-41, median 10); HS=12.34 days (5-88, median 10) 1) unpublished study; individual patient data obtained from author (JG Docherty), all patients included were randomised
Participants
Interventions
Outcomes
Notes
Risk of bias
Stapled versus handsewn methods for ileocolic anastomoses (Review) Copyright 2011 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 14
Docherty 1991
(Continued)
Bias
Authors judgement
Support for judgement Batch of sealed envelopes
Random sequence generation (selection Low risk bias) Allocation concealment (selection bias) Low risk
Outcome measurements are not likely to be inuenced by lack of blinding
Docherty 1995 Methods 1) Study location: ve surgical units in Scotland 2) Study duration: April 1985 to April 1989 3) Follow-up duration: at the discretion of surgeon 4) Randomisation technique: stratied randomisation - equal number of sealed envelopes indicating sutures or staples for each surgeon 5) Randomisation time: after resection when surgeon satised either technique feasible and equally appropriate 6) Blinding: not described 1) Source population: not described 2) Type of disease: no restriction 3) Type of anastomosis: oesophageal, upper gastrointestinal, colonic (ileocolic, colocolic, colostomy closures), colorectal 4) Inclusion criteria: all patients under the care of a participating surgeon undergoing elective or emergency surgery which was likely to results in a gastrointestinal anastomosis. 5) Exclusion criteria: in patients for whom one or another technique was considered to offer a particular advantage, randomisation did not take place. They were observed for follow-up in an identical way but analysed separately (ileocolic n=17) (p.178) 6) Number of participants: total n=1169, total randomised n=1004, ileocolic n=272, ileocolic randomised n=255 (cancer ileocolic n=190) 7) Excluded post randomisation: those who had palliative resection (n=97 altogether; ileocolic n=70) in the published study but they are included in individual patient data 8) Lost to follow-up: for cancer group incomplete follow-up n=15 altogether & not in nal study group (p.179) 1) Group S: combinations of GIA, TA and EEA, ileocolic n=133 (cancer n=102) 2) Group HS: either single layer interrupted 2/0 polyamide or in two layers with inner continuous 2/0 polyglycolic and outer 2/0 polyamide, ileocolic n=122 (cancer n=88) 1) Overall anastomotic leak: S=1 (0.8%), HS=4 (3.3%) 2) Clinical anastomotic leak: S=1 (0.8%), HS =4 (3.3%) 3) Anastomotic time: S=8.72 min (2-30, median 7); HS=22.36 min (8-90, median 20) 4) Operative mortality: S=6 (4.5%), HS=6 (4.9% ) 5) Intra-abdominal abscess: S=1 (0.8%), HS=4 (3.3%) 6) Wound infection: S=14 (10.5%), HS=9 (7.4%)
15
Participants
Interventions
Outcomes
Docherty 1995
(Continued)
7) Length of stay: S=13.63 days (2-30, median 7); HS=13.25 days (8-90, median 20) Notes 1) Randomisation success: the patients in sutured group versus stapled group is matched in terms of baseline characteristics 2) Representativity: situation was thought to reect average surgical practice in UK 3) Sample size: 1000 randomised patients gives adequate statistical power to detect a difference between true leak rates of 5% and 10% at a 5% signicance level 4) Intention to treat: yes 5) Learning curve: consultants had preliminary experience with surgical stapling, only one consultant is a regular stapler user, junior staff contributed, none of the surgeons has specic GI interest 6) In 10% random sample of study patients the key study variables were checked against original case notes as data recording audit. 7) Includes data published in West 1991, Akyol 1991 and individual data obtained from author (JG Docherty)
Risk of bias Bias Authors judgement Support for judgement Stratied randomisation
Sealed envelopes Outcome measurements are not likely to be inuenced by lack of blinding
Blinding (performance bias and detection Low risk bias) Subjective outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Other bias Ikeuchi 2000 Methods Low risk
No missing outcome data
Low risk Low risk
Reported all pre-specied outcomes
1) 1) Study location: Hyogo, Japan 2) Study duration: July 1987 to August 1996 3) Follow-up duration: median 87 (range 36-140) months for all groups 4) Randomisation technique: sealed envelopes (random fashion) 5) Randomisation time: during the operation after resection 6) Blinding: mentioned to be double-blind but not explained 1) Source population: patients with Crohns disease who underwent intestinal resection at the institution 2) Type of disease: Crohns disease
16
Participants
Ikeuchi 2000
(Continued)
3) Type of anastomosis: ileoileal, ileocolic, colocolic, ileorectal 4) Inclusion criteria: 68 patients with Crohns disease who underwent intestinal resection 5) Exclusion criteria: not described 6) Number of participants: ileocolic n=29 (30 anastomoses) (total in study, n=68) 7) Excluded post randomisation: none 8) Lost to follow-up: n=5 for whole study (p.494 geographical distance, death) 9) Age: 26.4 (15-44, median 25) 10) Sex: M=20, F=9 Interventions 1) Group S: functional end-to-end - linear stapling instrument (RL60), and linear anastomotic instrument (PLC50 or PLC75) (ileocolic n=11) 2) Group HS: layer-to-layer with inner running 3-0 Monocryl and outer interrupted 40 Vicryl (ileocolic n=18) (1 patient had ileocaecal resection then right hemicolectomy as reoperation, only ileocaecal resection included for outcomes) 1) Overall anastomotic leak: 0 2) Clinical anastomotic leak: 0 3) Anastomotic stricture: S=1, HS=6 4) Anastomotic haemorrhage: 0 5) Anastomotic time: individual patient data unavailable 6) Re-operation: S=1, HS=6 7) Operative mortality: 0 8) Intra-abdominal abscess: 0 9) Wound infection: S=0, HS=5 10) Length of stay: Stapled: 30 days (21-51, median 30); Handsewn: 23.7 days (16-50, median 21.5) 1) Randomisation success: no signicant difference between the two groups in patient characteristics 2) Representativity: not mentioned 3) Sample size: not calculated 4) Intention to treat: not mentioned 5) Learning curve: not assessed 6) Others: patients on elemental diet for at least 2 months post surgery 7) Individual data not available in publication were obtained from correspondence with author (H Ikeuchi) 8) Includes data published in Kusunoki 1998
Outcomes
Notes
Risk of bias Bias Authors judgement Support for judgement Random sequence generation
Sealed envelopes
17
Ikeuchi 2000
(Continued)
Blinding (performance bias and detection Low risk bias) Subjective outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Other bias Izbicki 1998 Methods Low risk
double-blind
Low risk Low risk
1) Study location: Hamburg 2) Study duration: 1 April 1993 to 31 December 1995 3) Follow-up duration: patients contacted by telephone at 3 months. Patients with complaints were invited for a personal visit. Data available up to 5 months. 4) Randomisation technique: not described 5) Randomisation time: after the resection if both handsewn and stapled anastomosis could be done 6) Blinding: not described 1) Source population: not described 2) Type of disease: carcinoma, non-cancer 3) Type of anastomosis: gastrectomy, gastric resection (Billroth II), Whipples procedure, segmental colonic resection, right hemicolectomy, left hemicolectomy, sigmoid- or anterior rectal resections, total colectomy with pouch-anal anastomosis 4) Inclusion criteria: patients with elective resection of the gastrointestinal tract; only included if both handsewn and stapled anastomosis could be done 5) Exclusion criteria: Crohns disease 6) Number of participants: right hemicolectomy n=36, total n=200 7) Excluded post randomisation: not mentioned 8) Lost to follow-up: n=9 (4.5%) 1) Group S: functional end-to-end with linear stapler (side-to-side ileo-transversostomy) (ileocolic n=15) 2) Group HS: single layer continuous (ileocolic n=21) 1) Overall anastomotic leak: S=1, HS=0 2) Clinical anastomotic leak: S=1, HS=0 3) Anastomotic stricture: 0 4) Anastomotic haemorrhage: 0 5) Operative mortality: 0 1) Randomisation success: difference in age between 2 groups not statistically signicant, no other baseline characteristics considered 2) Representativity: not mentioned 3) Sample size: not mentioned
18
Participants
Interventions
Outcomes
Notes
Izbicki 1998
(Continued)
4) Intention to treat: not mentioned 5) Learning curve: not mentioned 6) Others: original paper in German 7) Authors were contacted for further information but it has not yet been provided Risk of bias Bias Authors judgement Support for judgement Outcome measurements are not likely to be inuenced by lack of blinding
Blinding (performance bias and detection Low risk bias) Subjective outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Low risk
Reasons for missing outcome data unlikely to be related to true outcome Reported all pre-specied outcomes[
Low risk
Kracht 1993 Methods 1) Study location: 29 centres in France (at least 10 patients operated on in each centre, at least 4 enrolled per year per centre) 2) Study duration: 1981 through 1990 3) Follow-up duration: 30 days following discharge 4) Randomisation technique: unfolding the previously folded and stapled, opaque right corner of structured forms, bound together sequentially, under which the type of anastomosis to be performed was concealed. Allocation order was established by computergenerated randomisation. Random assignment was balanced within each centre and adjusted for each 5-10 patients. 5) Randomisation time: during operation once the colon had been resected and the surgeon was sure that all techniques were feasible and appropriate 6) Blinding: of outcome assessment (radiologist) 1) Source population: not described 2) Type of disease: right colonic adenocarcinoma 3) Type of anastomosis: ileocolic 4) Inclusion criteria: 457 consecutive patients undergoing right hemicolectomy for right colonic adenocarcinoma, whether with curative or palliative intent, no limits on age, general health, degree of tumour growth or extension, patients with contained pericolic abscess or infected tumour were included 5) Exclusion criteria: non-resection of the right colon, right hemicolectomy performed for inammatory or non-malignant lesions, as well as emergency resections for which preoperative colonic preparation was impossible, presence of overt local sepsis 6) Number of participants: n=457 7) Excluded post randomisation: n=17 (protocol violation: use of technique other than the one randomly assigned, denitive pathology examination showing a non-malignant lesion, or emergency resection), equally distributed between ve groups, no statistically signicant difference in assessment criteria among the withdrawn patients. 8) Lost to follow-up: none mentioned
19
Participants
Kracht 1993
(Continued)
9) Age: mean 70.2 years +/- 12.3 years (range 27 to 92 years) 10) Sex: F=261, M=179 Interventions 1) Group S: side-to-side, using GIA and TA-90 stapling devices, n=106 2) Group HS: n=334, all done with one extramucosal layer with polyglycolic acid or polygalactin 3/0 (1) end-to-end with interrupted sutures (EEI) n=84; (2) end-to-end with continuous sutures (EEC) n=77; (3) end-to-side with interrupted sutures (ESI) n= 82; (4) end-to-side with continuous sutures (ESC) colonic stump closed either manual or staples according to surgeon preference, n=91 1) Overall anastomotic leakage: S=3 (2.8%), HS=28 (8.4%) 2) Clinical anastomotic leak: S=2 (1.9%), HS=15 (4.5%) 3) Radiological anastomotic leak (at eighth to tenth postoperative day): S=1 (0.9%), HS=13 (3.9%) 4) Mortality: S=2 (1.9%), HS=12 (3.6%) 5) Intra-abdominal abscess or peritonitis: S=2 (1.9%), HS= 14 (4.2%) 6) Wound abscess: S=10 (9.4%), HS=31 (9.3%) 1) In total, Dukes: A=68; B=176; C=112 (all incomplete) 2) Randomisation success: preoperative criteria were comparable, except there were signicantly more women in stapled group. Comparable in pre-operative, intra-operative & pathological criteria except intra-operative septic spillage was signicantly lower in stapled group; no difference between centres concerning patient demographics, colonic preparation, or outcome; comparable regarding curative (78%) or palliative (22%) nature of operation, tumour spread (Dukes); no statistically signicant difference in number of stumps closed by hand or staples in end-to-side technique 3) Representativity: protocol constructed to correspond to the practical aspects of everyday surgery, 457 consecutive patients 4) Sample size: designed as a pragmatic trial as dened by Schwartz. The hypothesis that a 5% reduction of anastomotic leakage (10-5%) could be obtained by a specic type of anastomosis, with an expected gamma error of 5%. The gamma error accounts for the risk in assuming that one treatment is better than another, even though it may be worse. Accordingly, the number of patients necessary for this study was calculated to be 420. More patients (100 patients) were included in the stapled group so that this group could be compared with the 4 sutured groups taken as a whole or individually (80 patients). 5) Intention to treat: not mentioned, not done as no lost to follow-up was mentioned 6) Learning curve: all surgeons performing anastomosis or preceptors assisting residents in these anastomoses had at least four years experience with both techniques 7) Schwartz D, Flamant R, Lellouch J (1980). Clinical trials. Academic Press Inc, London. 8) Same data also published in Kracht 1990
Outcomes
Notes
Risk of bias Bias Authors judgement Support for judgement Computer generated randomisation
Random sequence generation (selection Low risk bias)
20
Kracht 1993
(Continued)
Allocation concealment (selection bias)
Low risk
Sequential sealed forms Radiologist who assessed radiological anastomotic leak
Blinding (performance bias and detection Low risk bias) Subjective outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Other bias McLeod 2009 Methods Low risk
Low risk Low risk
1) Study location: 10 Canadian, 6 American, 1 British centres 2) Study duration: Jan 2001 to Jul 2004 3) Follow-up duration: 12 months 4) Randomisation technique: centralised computer generated block randomisation within each strata stratied by centre 5) Randomisation time: intraoperatively when either type of anastomosis could be performed safely 6) Blinding: patient, gastroenterologist who performed the colonoscopy to assess disease recurrence 1) Source population: patients were accrued from ten Canadian, six American, and one British centre 2) Type of disease: Crohns disease 3) Type of anastomosis: ileocolic. Surgery was performed open or laparoscopic depending on patient factors and surgeon preference 4) Inclusion criteria: patients with Crohns disease limited to the distal ileum and right colon who were scheduled for an elective ileocolic resection, patients with an internal stula who required resection of an otherwise normal segment of bowel, patients with minimal perianal disease that did not require treatment 5) Exclusion criteria: previous resection, requirement for a defunctioning ileostomy, on medications for Crohns disease that could not be discontinued postoperatively 6) Number of participants: n=171 7) Excluded post randomisation: n=1 (had terminal ileal resection rather than ileocolic resection) 8) Lost to follow-up: total n=31: Lost to follow-up n=9 (S=6, HS=3), n=22 (S=12, HS= 10) refused follow-up investigations despite being followed for a mean of 11.2 +/- 1. 9 months. The follow-up investigation was a colonoscopy to check for recurrence.) If symptomatic recurrence within the 12 months follow up period, the participation in the study was terminated as they were considered to have reached an end point 9) Age: mean age S=40.3 years, HS=38.2 years 10) Sex: male to female ratio of 30:54 for stapled group and 32:54 for handsewn group 11) Notes: change of study protocol. Initially used 5-ASA maintenance therapy, then no maintenance therapy as 5-ASA was not approved. After 72 patients were entered into
21
Participants
McLeod 2009
(Continued)
the trial, due to problems with accrual an investigators meeting was held in November 2002. It was decided to include individuals who t the inclusion criteria but were advised to take postoperative azathioprine. The decision regarding the need for maintenance therapy had to be made pre-operatively, and patients were stratied to an azathioprine stratum or no maintenance therapy stratum before they were randomised to the two treatment groups Interventions 1) Group S: stapled side-to-side anastomosis (STSA), either side-to-side with stapler passing through enterotomies, or functional end-to-end, both using a TLC 100 mm stapler or a transverse stapler, n=84 2) Group HS: end-to-end anastomosis (ETEA), either a single-layer or two-layer anastomosis with 2-0 PDS, n=86 1) Clinical anastomotic leak: S=6 (7%), HS=6 (7%), p=0.86 2) Anastomotic time (range not reported): S=15 minutes, HS=31 minutes (p<0.0001) 3) Re-operation: S=6 (7%), HS=6 (7%), p=0.86 4) Operative mortality: S=0, HS=0 5) Intra-abdominal abscess: S=1 (1%), HS=1 (1%) 6) Wound infection: S=9 (11%), HS=8 (9%) 7) Length of hospital stay (range not reported): S=8.3 days, HS=6.8 days (p<0.05) 1) Randomisation success: baseline characteristics were similar. However there were more females compared with males in the whole trial 2) Representativity: not mentioned 3) Sample size: The main outcome of the study was endoscopic recurrence. The statistical power of the study was 70 percent to detect a signicant absolute risk difference of 20 percent, or 80 percent power to detect an absolute risk difference of 23 percent 4) Intention to treat: no. Although none of the patients lost to follow-up or those who refused follow-up colonoscopy had symptoms suggestive of recurrent disease, they were not included in the analysis of disease recurrence 5) Learning curve: all surgery was performed by colorectal surgeons experienced in the surgical treatment of Crohns disease
Outcomes
Notes
Risk of bias Bias Authors judgement Support for judgement The study used computer generated block randomisation stratied by centre The study used a central phone-in randomisation service, which presumably kept the allocation concealed To minimize observer bias, where possible, a gastroenterologist performed the colonoscopy
22
McLeod 2009
(Continued)
Incomplete outcome data (attrition bias) All outcomes
Low risk
Patients who were lost to follow up (6/84 in stapled group, 3/86 in handsewn group) or who refused follow up investigation (12/ 84 in stapled group, 10/86 in handsewn group) were explained It is clear that the published report included pre-specied and expected outcomes
Selective reporting (reporting bias)
Low risk
Other bias
Low risk
Characteristics of excluded studies [ordered by study ID]
Study Alessandroni 2009 Anwar 2004 Brennan 1982 Brundage 1999 Bubrick 1991 Cajozzo 1990 Chassin 1978 Corman 1989 Demetriades 2002 Dyess 1990 Hashemi 1998 Munoz-Juarez 2001 Reiling 1980 Resegotti 2005 Scarpa 2004
Reason for exclusion not randomised not randomised circular stapler not randomised not randomised between stapled and handsewn anastomosis EEA not function end-to-end non-randomised retrospective comparison not randomised between stapled and handsewn anastomosis not randomised not randomised between stapled and handsewn anastomosis longitudinal study, not randomised not RCT, is case-control not functional end to end techniques chosen by surgeons preference. The randomised proportion is not analysed separately not randomised
23
(Continued)
Scarpa 2007 Scher 1982 Tersigni 2003 Wolmark 1986 Yamamoto 1999
not randomised not randomised not randomised not randomised retrospective non-randomised comparison
24
DATA AND ANALYSES
Comparison 1. All studies
Outcome or subgroup title 1 Overall anastomotic leak 2 Clinical anastomotic leak 3 Radiological anastomotic leak 4 Anastomotic stricture 5 Anastomotic haemorrhage 6 Anastomotic time (min) 7 Re-operation 8 Operative mortality 9 Intra-abdominal abscess 10 Wound infection 11 Length of hospital stay (day)
No. of studies 7 7 1 2 2 1 2 6 5 5 3
No. of participants 1125 1125 440 65 65 255 199 1087 932 932 424
Statistical method Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
Effect size 0.48 [0.24, 0.95] 0.55 [0.27, 1.15] 0.24 [0.03, 1.82] 0.2 [0.02, 1.95] 0.0 [0.0, 0.0] -13.64 [-15.86, -11. 42] 0.67 [0.25, 1.81] 0.74 [0.33, 1.65] 0.41 [0.14, 1.25] 1.05 [0.66, 1.70] 0.19 [-1.50, 1.87]
Comparison 2. Cancer
Outcome or subgroup title 1 Overall anastomotic leak 2 Clinical anastomotic leak 3 Radiological anastomotic leak 4 Anastomotic time 5 Operative mortality 6 Intra-abdominal abscess 7 Wound infection 8 Length of stay
No. of studies 4 4 1 1 3 3 3 2
No. of participants 825 825 440 190 787 668 668 334
Statistical method Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
Effect size 0.28 [0.10, 0.75] 0.30 [0.10, 0.95] 0.24 [0.03, 1.82] -13.64 [-16.26, -11. 02] 0.57 [0.23, 1.41] 0.39 [0.11, 1.37] 1.19 [0.66, 2.14] -0.42 [-2.30, 1.45]
Comparison 3. Non-Cancer
Outcome or subgroup title 1 Overall anastomotic leak 2 Clinical anastomotic leak 3 Anastomotic stricture 4 Anastomotic haemorrhage
No. of studies 3 3 1 1
No. of participants 264 264 29 29
Statistical method Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI)
Effect size 0.89 [0.30, 2.64] 0.89 [0.30, 2.64] 0.2 [0.02, 1.95] 0.0 [0.0, 0.0]
25
5 Anastomotic time 6 Re-operation 7 Operative mortality 8 Intra-abdominal abscess 9 Wound infection 10 Length of stay
1 2 3 3 3 2
65 199 264 264 264 90
Mean Difference (IV, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
-14.27 [-18.54, -8. 00] 0.67 [0.25, 1.81] 3.54 [0.35, 35.93] 0.63 [0.08, 4.86] 0.83 [0.37, 1.86] 3.12 [-0.78, 7.03]
Analysis 1.1. Comparison 1 All studies, Outcome 1 Overall anastomotic leak.

Review: Stapled versus handsewn methods for ileocolic anastomoses
Comparison: 1 All studies Outcome: 1 Overall anastomotic leak
Study or subgroup
Stapled n/N
Handsewn n/N 0/16 4/87 4/122 0/18 0/21 28/334 6/86
Odds Ratio M-H,Fixed,95% CI
Odds Ratio M-H,Fixed,95% CI 0.0 [ 0.0, 0.0 ] 0.13 [ 0.01, 2.49 ] 0.22 [ 0.02, 2.03 ] 0.0 [ 0.0, 0.0 ] 4.45 [ 0.17, 116.94 ] 0.32 [ 0.09, 1.07 ] 1.03 [ 0.32, 3.32 ]
Didolkar 1986 Docherty 1991 Docherty 1995 Ikeuchi 2000 Izbicki 1998 Kracht 1993 McLeod 2009
0/22 0/70 1/133 0/11 1/15 3/106 6/84
Total (95% CI)
441
684
0.48 [ 0.24, 0.95 ]
Total events: 11 (Stapled), 42 (Handsewn) Heterogeneity: Chi2 = 5.04, df = 4 (P = 0.28); I2 =21% Test for overall effect: Z = 2.12 (P = 0.034) Test for subgroup differences: Not applicable
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
26
Analysis 1.2. Comparison 1 All studies, Outcome 2 Clinical anastomotic leak.

Comparison: 1 All studies Outcome: 2 Clinical anastomotic leak
Study or subgroup
Stapled n/N
Handsewn n/N 0/16 4/87 4/122 0/18 0/21 15/334 6/86
Odds Ratio M-H,Fixed,95% CI 0.0 [ 0.0, 0.0 ] 0.13 [ 0.01, 2.49 ] 0.22 [ 0.02, 2.03 ] 0.0 [ 0.0, 0.0 ] 4.45 [ 0.17, 116.94 ] 0.41 [ 0.09, 1.82 ] 1.03 [ 0.32, 3.32 ]
Didolkar 1986 Docherty 1991 Docherty 1995 Ikeuchi 2000 Izbicki 1998 Kracht 1993 McLeod 2009
0/22 0/70 1/133 0/11 1/15 2/106 6/84
Total (95% CI)
441
684
0.55 [ 0.27, 1.15 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
27
Analysis 1.3. Comparison 1 All studies, Outcome 3 Radiological anastomotic leak.

Comparison: 1 All studies Outcome: 3 Radiological anastomotic leak
Study or subgroup
Stapled n/N
Handsewn n/N 13/334
Weight
Kracht 1993
1/106
100.0 %
0.24 [ 0.03, 1.82 ]
Total (95% CI)

Heterogeneity: not applicable
106
334
100.0 %
0.24 [ 0.03, 1.82 ]
Total events: 1 (Stapled), 13 (Handsewn) Test for overall effect: Z = 1.39 (P = 0.17) Test for subgroup differences: Not applicable
0.1 0.2
0.5
10
Favour stapled
Favours handsewn
Analysis 1.4. Comparison 1 All studies, Outcome 4 Anastomotic stricture.

Comparison: 1 All studies Outcome: 4 Anastomotic stricture
Study or subgroup
Stapled n/N
Handsewn n/N 6/18 0/21
Odds Ratio M-H,Fixed,95% CI 0.20 [ 0.02, 1.95 ] 0.0 [ 0.0, 0.0 ]
Ikeuchi 2000 Izbicki 1998
1/11 0/15
Total (95% CI)

Total events: 1 (Stapled), 6 (Handsewn)
26
39
0.20 [ 0.02, 1.95 ]
Heterogeneity: Chi2 = 0.0, df = 0 (P = 1.00); I2 =0.0% Test for overall effect: Z = 1.39 (P = 0.17) Test for subgroup differences: Not applicable
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
28
Analysis 1.5. Comparison 1 All studies, Outcome 5 Anastomotic haemorrhage.

Comparison: 1 All studies Outcome: 5 Anastomotic haemorrhage
Study or subgroup
Stapled n/N
Odds Ratio M-H,Fixed,95% CI 0.0 [ 0.0, 0.0 ] 0.0 [ 0.0, 0.0 ]
Ikeuchi 2000 Izbicki 1998
0/11 0/15
Total (95% CI)

26
39
0.0 [ 0.0, 0.0 ]
Heterogeneity: Chi2 = 0.0, df = 0 (P<0.00001); I2 =0.0% Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
Analysis 1.6. Comparison 1 All studies, Outcome 6 Anastomotic time (min).

Comparison: 1 All studies Outcome: 6 Anastomotic time (min)
Study or subgroup
Stapled N Mean(SD) 8.72 (5.12)
Handsewn N 122 Mean(SD) 22.36 (11.54)
Mean Difference IV,Fixed,95% CI
Weight
Docherty 1995
133
100.0 %
-13.64 [ -15.86, -11.42 ]
Total (95% CI)
133
122
100.0 %
-13.64 [ -15.86, -11.42 ]
Heterogeneity: not applicable Test for overall effect: Z = 12.02 (P < 0.00001) Test for subgroup differences: Not applicable
-10
-5
10
Favours stapled
Favours handsewn
29
Analysis 1.7. Comparison 1 All studies, Outcome 7 Re-operation.

Comparison: 1 All studies Outcome: 7 Re-operation
Study or subgroup
Stapled n/N
Weight
Ikeuchi 2000 McLeod 2009
1/11 6/84
42.9 % 57.1 %
0.20 [ 0.02, 1.95 ] 1.03 [ 0.32, 3.32 ]
Total (95% CI)
95
104
100.0 %
0.67 [ 0.25, 1.81 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
30
Analysis 1.8. Comparison 1 All studies, Outcome 8 Operative mortality.

Comparison: 1 All studies Outcome: 8 Operative mortality
Study or subgroup
Stapled n/N
Handsewn n/N 3/87 6/122 0/18 0/21 12/334 0/86
Odds Ratio M-H,Fixed,95% CI 0.82 [ 0.13, 5.07 ] 0.91 [ 0.29, 2.91 ] 0.0 [ 0.0, 0.0 ] 0.0 [ 0.0, 0.0 ] 0.52 [ 0.11, 2.34 ] 0.0 [ 0.0, 0.0 ]
Docherty 1991 Docherty 1995 Ikeuchi 2000 Izbicki 1998 Kracht 1993 McLeod 2009
2/70 6/133 0/11 0/15 2/106 0/84
Total (95% CI)
419
668
0.74 [ 0.33, 1.65 ]
Total events: 10 (Stapled), 21 (Handsewn) Heterogeneity: Chi2 = 0.36, df = 2 (P = 0.84); I2 =0.0% Test for overall effect: Z = 0.74 (P = 0.46) Test for subgroup differences: Not applicable
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
31
Analysis 1.9. Comparison 1 All studies, Outcome 9 Intra-abdominal abscess.

Comparison: 1 All studies Outcome: 9 Intra-abdominal abscess Study or subgroup Stapled n/N Didolkar 1986 Docherty 1995 Ikeuchi 2000 Kracht 1993 McLeod 2009 0/22 1/133 0/11 2/106 1/84 Handsewn n/N 0/16 4/122 0/18 14/334 1/86 Odds Ratio M-H,Fixed,95% CI Odds Ratio M-H,Fixed,95% CI 0.0 [ 0.0, 0.0 ] 0.22 [ 0.02, 2.03 ] 0.0 [ 0.0, 0.0 ] 0.44 [ 0.10, 1.97 ] 1.02 [ 0.06, 16.64 ]
Total (95% CI)
356
576
0.41 [ 0.14, 1.25 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
32
Analysis 1.10. Comparison 1 All studies, Outcome 10 Wound infection.

Comparison: 1 All studies Outcome: 10 Wound infection Study or subgroup Stapled n/N Didolkar 1986 Docherty 1995 Ikeuchi 2000 Kracht 1993 McLeod 2009 0/22 14/133 0/11 10/106 9/84 Handsewn n/N 0/16 9/122 5/18 31/334 8/86 Odds Ratio M-H,Fixed,95% CI Odds Ratio M-H,Fixed,95% CI 0.0 [ 0.0, 0.0 ] 1.48 [ 0.62, 3.55 ] 0.11 [ 0.01, 2.14 ] 1.02 [ 0.48, 2.15 ] 1.17 [ 0.43, 3.19 ]
Total (95% CI)
356
576
1.05 [ 0.66, 1.70 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
33
Analysis 1.11. Comparison 1 All studies, Outcome 11 Length of hospital stay (day).
Comparison: 1 All studies Outcome: 11 Length of hospital stay (day)
Study or subgroup
Stapled N Mean(SD) 11.36 (5.87) 13.63 (10.93) 30 (8.44)
Handsewn N 84 116 18 Mean(SD) 12.34 (9.46) 13.25 (8.89) 23.7 (7.76)
Weight
Docherty 1991 Docherty 1995 Ikeuchi 2000
68 127 11
47.0 % 45.5 % 7.5 %
-0.98 [ -3.44, 1.48 ] 0.38 [ -2.12, 2.88 ] 6.30 [ 0.16, 12.44 ]
Total (95% CI)
206
218
100.0 %
0.19 [ -1.50, 1.87 ]
Heterogeneity: Chi2 = 4.69, df = 2 (P = 0.10); I2 =57% Test for overall effect: Z = 0.22 (P = 0.83) Test for subgroup differences: Not applicable
-10
-5
10
Favours stapled
Favours handsewn
34
Analysis 2.1. Comparison 2 Cancer, Outcome 1 Overall anastomotic leak.

Comparison: 2 Cancer Outcome: 1 Overall anastomotic leak
Study or subgroup
Stapled n/N
Handsewn n/N 0/16 4/87 3/88 28/334
Odds Ratio M-H,Fixed,95% CI 0.0 [ 0.0, 0.0 ] 0.13 [ 0.01, 2.49 ] 0.28 [ 0.03, 2.75 ] 0.32 [ 0.09, 1.07 ]
Didolkar 1986 Docherty 1991 Docherty 1995 Kracht 1993
0/22 0/70 1/102 3/106
Total (95% CI)
300
525
0.28 [ 0.10, 0.75 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
35
Analysis 2.2. Comparison 2 Cancer, Outcome 2 Clinical anastomotic leak.

Comparison: 2 Cancer Outcome: 2 Clinical anastomotic leak
Study or subgroup
Stapled n/N
Handsewn n/N 0/16 4/87 3/88 15/334
Odds Ratio M-H,Fixed,95% CI 0.0 [ 0.0, 0.0 ] 0.13 [ 0.01, 2.49 ] 0.28 [ 0.03, 2.75 ] 0.41 [ 0.09, 1.82 ]
Didolkar 1986 Docherty 1991 Docherty 1995 Kracht 1993
0/22 0/70 1/102 2/106
Total (95% CI)
300
525
0.30 [ 0.10, 0.95 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
Analysis 2.3. Comparison 2 Cancer, Outcome 3 Radiological anastomotic leak.

Comparison: 2 Cancer Outcome: 3 Radiological anastomotic leak
Study or subgroup
Stapled n/N
Handsewn n/N 13/334
Weight
Kracht 1993
1/106
100.0 %
0.24 [ 0.03, 1.82 ]
Total (95% CI)

106
334
100.0 %
0.24 [ 0.03, 1.82 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
36
Analysis 2.4. Comparison 2 Cancer, Outcome 4 Anastomotic time.

Comparison: 2 Cancer Outcome: 4 Anastomotic time
Study or subgroup
Weight
Docherty 1995
102
100.0 %
-13.64 [ -16.26, -11.02 ]
Total (95% CI)
102
88
100.0 %
-13.64 [ -16.26, -11.02 ]
-10
-5
10
Favours stapled
Favours handsewn
37
Analysis 2.5. Comparison 2 Cancer, Outcome 5 Operative mortality.

Comparison: 2 Cancer Outcome: 5 Operative mortality
Study or subgroup
Stapled n/N
Handsewn n/N 3/87 5/88 12/334
Weight
Docherty 1991 Docherty 1995 Izbicki 1998
2/70 3/102 2/106
19.3 % 38.6 % 42.1 %
0.82 [ 0.13, 5.07 ] 0.50 [ 0.12, 2.17 ] 0.52 [ 0.11, 2.34 ]
Total (95% CI)
278
509
100.0 %
0.57 [ 0.23, 1.41 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
Analysis 2.6. Comparison 2 Cancer, Outcome 6 Intra-abdominal abscess.

Comparison: 2 Cancer Outcome: 6 Intra-abdominal abscess
Study or subgroup
Stapled n/N
Handsewn n/N 0/16 3/88 14/334
Odds Ratio M-H,Fixed,95% CI 0.0 [ 0.0, 0.0 ] 0.28 [ 0.03, 2.75 ] 0.44 [ 0.10, 1.97 ]
Didolkar 1986 Docherty 1995 Kracht 1993
0/22 1/102 2/106
Total (95% CI)
230
438
0.39 [ 0.11, 1.37 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
38
Analysis 2.7. Comparison 2 Cancer, Outcome 7 Wound infection.

Comparison: 2 Cancer Outcome: 7 Wound infection
Study or subgroup
Stapled n/N
Handsewn n/N 0/16 7/88 31/334
Didolkar 1986 Docherty 1995 Kracht 1993
0/22 12/102 10/106
Total (95% CI)
230
438
1.19 [ 0.66, 2.14 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
39
Analysis 2.8. Comparison 2 Cancer, Outcome 8 Length of stay.

Comparison: 2 Cancer Outcome: 8 Length of stay
Study or subgroup
Stapled N Mean(SD) 11.36 (5.87) 13.63 (10.93)
Handsewn N 84 83 Mean(SD) 12.34 (9.46) 13.29 (8.92)
Weight
Docherty 1991 Docherty 1995
68 99
57.9 % 42.1 %
-0.98 [ -3.44, 1.48 ] 0.34 [ -2.54, 3.22 ]
Total (95% CI)
167
167
100.0 %
-0.42 [ -2.30, 1.45 ]
-10
-5
10
Favours stapled
Favours handsewn
Analysis 3.1. Comparison 3 Non-Cancer, Outcome 1 Overall anastomotic leak.

Comparison: 3 Non-Cancer Outcome: 1 Overall anastomotic leak
Study or subgroup
Stapled n/N
Handsewn n/N 1/34 0/18 6/86
Docherty 1995 Ikeuchi 2000 McLeod 2009
0/31 0/11 6/84
Total (95% CI)

126
138
0.89 [ 0.30, 2.64 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
40
Analysis 3.2. Comparison 3 Non-Cancer, Outcome 2 Clinical anastomotic leak.

Comparison: 3 Non-Cancer Outcome: 2 Clinical anastomotic leak
Study or subgroup
Stapled n/N
Handsewn n/N 1/34 0/18 6/86
0/31 0/11 6/84
Total (95% CI)

126
138
0.89 [ 0.30, 2.64 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
41
Analysis 3.3. Comparison 3 Non-Cancer, Outcome 3 Anastomotic stricture.

Comparison: 3 Non-Cancer Outcome: 3 Anastomotic stricture
Study or subgroup
Stapled n/N
Handsewn n/N 6/18
Weight
Ikeuchi 2000
1/11
100.0 %
0.20 [ 0.02, 1.95 ]
Total (95% CI)

11
18
100.0 %
0.20 [ 0.02, 1.95 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
Analysis 3.4. Comparison 3 Non-Cancer, Outcome 4 Anastomotic haemorrhage.

Comparison: 3 Non-Cancer Outcome: 4 Anastomotic haemorrhage
Study or subgroup
Stapled n/N
Handsewn n/N 0/18
Odds Ratio M-H,Fixed,95% CI 0.0 [ 0.0, 0.0 ]
Ikeuchi 2000
0/11
Total (95% CI)

Total events: 0 (Stapled), 0 (Handsewn) Heterogeneity: not applicable
11
18
0.0 [ 0.0, 0.0 ]
Test for overall effect: Z = 0.0 (P < 0.00001) Test for subgroup differences: Not applicable
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
42
Analysis 3.5. Comparison 3 Non-Cancer, Outcome 5 Anastomotic time.

Comparison: 3 Non-Cancer Outcome: 5 Anastomotic time
Study or subgroup
Weight
Docherty 1995
31
100.0 %
-14.27 [ -18.54, -10.00 ]
Total (95% CI)
31
34
100.0 %
-14.27 [ -18.54, -10.00 ]
-10
-5
10
Favours stapled
Favours handsewn
Analysis 3.6. Comparison 3 Non-Cancer, Outcome 6 Re-operation.

Comparison: 3 Non-Cancer Outcome: 6 Re-operation
Study or subgroup
Stapled n/N
Weight
Ikeuchi 2000 McLeod 2009
1/11 6/84
42.9 % 57.1 %
0.20 [ 0.02, 1.95 ] 1.03 [ 0.32, 3.32 ]
Total (95% CI)
95
104
100.0 %
0.67 [ 0.25, 1.81 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
43
Analysis 3.7. Comparison 3 Non-Cancer, Outcome 7 Operative mortality.

Comparison: 3 Non-Cancer Outcome: 7 Operative mortality
Study or subgroup
Stapled n/N
Handsewn n/N 1/34 0/18 0/86
3/31 0/11 0/84
Total (95% CI)

126
138
3.54 [ 0.35, 35.93 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
Analysis 3.8. Comparison 3 Non-Cancer, Outcome 8 Intra-abdominal abscess.

Comparison: 3 Non-Cancer Outcome: 8 Intra-abdominal abscess
Study or subgroup
Stapled n/N
Handsewn n/N 1/34 0/18 1/86
0/31 0/11 1/84
Total (95% CI)

126
138
0.63 [ 0.08, 4.86 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
44
Analysis 3.9. Comparison 3 Non-Cancer, Outcome 9 Wound infection.

Comparison: 3 Non-Cancer Outcome: 9 Wound infection
Study or subgroup
Stapled n/N
Handsewn n/N 2/34 5/18 8/86
Weight
2/31 0/11 9/84
13.8 % 31.6 % 54.6 %
1.10 [ 0.15, 8.35 ] 0.11 [ 0.01, 2.14 ] 1.17 [ 0.43, 3.19 ]
Total (95% CI)
126
138
100.0 %
0.83 [ 0.37, 1.86 ]
0.1 0.2
0.5
10
Favours stapled
Favours handsewn
45
Analysis 3.10. Comparison 3 Non-Cancer, Outcome 10 Length of stay.

Comparison: 3 Non-Cancer Outcome: 10 Length of stay
Study or subgroup
Stapled N Mean(SD) 13.84 (11.21) 30 (8.44)
Handsewn N 33 18 Mean(SD) 12.87 (8.45) 23.7 (7.76)
Weight
Docherty 1995 Ikeuchi 2000
28 11
59.6 % 40.4 %
0.97 [ -4.08, 6.02 ] 6.30 [ 0.16, 12.44 ]
Total (95% CI)
39
51
100.0 %
3.12 [ -0.78, 7.03 ]
Heterogeneity: Chi2 = 1.72, df = 1 (P = 0.19); I2 =42% Test for overall effect: Z = 1.57 (P = 0.12) Test for subgroup differences: Not applicable
-10
-5
10
Favours stapled
Favours handsewn
WHATS NEW
Last assessed as up-to-date: 23 April 2011.
Date 24 April 2011
Event
Description
New citation required but conclusions have not changed added a new author & one included study, updated RoB tables & text New search has been performed 1st update
24 April 2011
HISTORY
Protocol rst published: Issue 3, 2003 Review rst published: Issue 3, 2007
46
Date 1 December 2010 5 August 2008 1 March 2007
Event New search has been performed Amended New citation required and conclusions have changed
Description Seach for new studies Converted to new review format. Substantive amendment
CONTRIBUTIONS OF AUTHORS
None mentioned
DECLARATIONS OF INTEREST
JG Docherty: Research had been funded by both Ethicon and by Autosuture (now Tyco). Both had helped with attending and presenting at scientic and educational meetings.
SOURCES OF SUPPORT Internal sources

Auckland Medical Research Foundation Summer Studentship, New Zealand.
External sources
No sources of support supplied
INDEX TERMS Medical Subject Headings (MeSH)

Surgical
Stapling [adverse effects]; Suture Techniques [adverse effects]; Anastomosis, Surgical [adverse effects; methods]; Colon
[ surgery]; Colorectal Neoplasms [surgery]; Crohn Disease [surgery]; Ileum [ surgery]; Randomized Controlled Trials as Topic; Surgical Wound Dehiscence [etiology]
47
MeSH check words

Adult; Humans
48

Stapled Versus Handsewn Methods For Ileocolic Anastomoses (Review)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Stapled Versus Handsewn Methods For Ileocolic Anastomoses (Review)

Uploaded by

Copyright:

Available Formats

Stapled versus handsewn methods for ileocolic anastomoses (Review)

SOURCES OF SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . INDEX TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Stapled versus handsewn methods for ileocolic anastomoses

Data collection and analysis

Risk of bias in included studies

AUTHORS CONCLUSIONS Implications for practice

Implications for research

References to studies included in this review

References to studies excluded from this review

CHARACTERISTICS OF STUDIES Characteristics of included studies [ordered by study ID]

Support for judgement Batch of sealed envelopes

Outcome measurements are not likely to be inuenced by lack of blinding

No missing outcome data

Low risk Low risk

Reported all pre-specied outcomes

No missing outcome data

Low risk Low risk

Reported all pre-specied outcomes

Random sequence generation (selection Low risk bias)

Allocation concealment (selection bias)

Sequential sealed forms Radiologist who assessed radiological anastomotic leak

No missing outcome data

Low risk Low risk

Reported all pre-specied outcomes

Incomplete outcome data (attrition bias) All outcomes

Selective reporting (reporting bias)

Characteristics of excluded studies [ordered by study ID]

DATA AND ANALYSES

Comparison 1. All studies

No. of participants 264 264 29 29

65 199 264 264 264 90

Analysis 1.1. Comparison 1 All studies, Outcome 1 Overall anastomotic leak.

Comparison: 1 All studies Outcome: 1 Overall anastomotic leak

Handsewn n/N 0/16 4/87 4/122 0/18 0/21 28/334 6/86

Odds Ratio M-H,Fixed,95% CI

0/22 0/70 1/133 0/11 1/15 3/106 6/84

Total (95% CI)

0.48 [ 0.24, 0.95 ]

Analysis 1.2. Comparison 1 All studies, Outcome 2 Clinical anastomotic leak.

Comparison: 1 All studies Outcome: 2 Clinical anastomotic leak

Handsewn n/N 0/16 4/87 4/122 0/18 0/21 15/334 6/86

Odds Ratio M-H,Fixed,95% CI

0/22 0/70 1/133 0/11 1/15 2/106 6/84

Total (95% CI)

0.55 [ 0.27, 1.15 ]

Analysis 1.3. Comparison 1 All studies, Outcome 3 Radiological anastomotic leak.

Comparison: 1 All studies Outcome: 3 Radiological anastomotic leak

Handsewn n/N 13/334

Odds Ratio M-H,Fixed,95% CI

Odds Ratio M-H,Fixed,95% CI

0.24 [ 0.03, 1.82 ]

Total (95% CI)

0.24 [ 0.03, 1.82 ]

Analysis 1.4. Comparison 1 All studies, Outcome 4 Anastomotic stricture.

Comparison: 1 All studies Outcome: 4 Anastomotic stricture

Handsewn n/N 6/18 0/21

Odds Ratio M-H,Fixed,95% CI

Odds Ratio M-H,Fixed,95% CI 0.20 [ 0.02, 1.95 ] 0.0 [ 0.0, 0.0 ]

Ikeuchi 2000 Izbicki 1998

Total (95% CI)

0.20 [ 0.02, 1.95 ]

Analysis 1.5. Comparison 1 All studies, Outcome 5 Anastomotic haemorrhage.