Cochrane Database of Systematic Reviews (Reviews) || Single layer versus double layer suture anastomosis of the gastrointestinal tract

Single layer versus double layer suture anastomosis of the

gastrointestinal tract (Review)

Sajid MS, Siddiqui MRS, Baig MK

This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library

2012, Issue 1

http://www.thecochranelibrary.com

Single layer versus double layer suture anastomosis of the gastrointestinal tract (Review)

Copyright © 2012 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.

http://www.thecochranelibrary.com

T A B L E O F C O N T E N T S

1HEADER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2PLAIN LANGUAGE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2SUMMARY OF FINDINGS FOR THE MAIN COMPARISON . . . . . . . . . . . . . . . . . . .

6BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Figure 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Figure 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

9OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

9METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

12RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Figure 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Figure 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Figure 7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Figure 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Figure 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Figure 10. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Figure 11. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

18DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19AUTHORS’ CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22CHARACTERISTICS OF STUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

33DATA AND ANALYSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Analysis 1.1. Comparison 1 Anastomosis, Outcome 1 Anastomosis leak. . . . . . . . . . . . . . . . . 33

Analysis 1.2. Comparison 1 Anastomosis, Outcome 2 Anastomosis time. . . . . . . . . . . . . . . . 34

Analysis 1.3. Comparison 1 Anastomosis, Outcome 3 Hospital stay. . . . . . . . . . . . . . . . . . 34

Analysis 1.4. Comparison 1 Anastomosis, Outcome 4 Mortality. . . . . . . . . . . . . . . . . . . 35

Analysis 1.5. Comparison 1 Anastomosis, Outcome 5 Major complications. . . . . . . . . . . . . . . 36

Analysis 1.6. Comparison 1 Anastomosis, Outcome 6 Anatomotic leak in high quality trials. . . . . . . . . . 37

Analysis 1.7. Comparison 1 Anastomosis, Outcome 7 Anastomotic leak in poor quality trials. . . . . . . . . 37

37APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39CONTRIBUTIONS OF AUTHORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39DECLARATIONS OF INTEREST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

40SOURCES OF SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

40DIFFERENCES BETWEEN PROTOCOL AND REVIEW . . . . . . . . . . . . . . . . . . . . .

40NOTES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

40INDEX TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

iSingle layer versus double layer suture anastomosis of the gastrointestinal tract (Review)


[Intervention Review]

Single layer versus double layer suture anastomosis of thegastrointestinal tract

Muhammad S Sajid1, Muhammed Rafay Sameem Siddiqui2 , Mirza K Baig1

1Department of Colorectal Surgery, Worthing Hospital, Worthing, UK. 2Department of Colorectal Surgery, St Marks Hospital, Harrow,

UK

Contact address: Muhammad S Sajid, Department of Colorectal Surgery, Worthing Hospital, Western Sussex Hospitals NHS Trust,

Worthing, West Sussex, BN11 2DH, UK. [email protected].

Editorial group: Cochrane Colorectal Cancer Group.

Publication status and date: New, published in Issue 1, 2012.

Review content assessed as up-to-date: 13 May 2011.

Citation: Sajid MS, Siddiqui MRS, Baig MK. Single layer versus double layer suture anastomosis of the gastrointestinal tract. Cochrane

Database of Systematic Reviews 2012, Issue 1. Art. No.: CD005477. DOI: 10.1002/14651858.CD005477.pub4.


A B S T R A C T

Background

Gastrointestinal anastomosis (GIA) is an essential step to maintain the continuity of gastrointestinal tract following intestinal resection.

GIA is still a source of significant controversy among surgeons due to the use of variety of approaches. Adequate apposition by single

layer or double layer anastomosis may affect outcome after GIA

Objectives

The objective of this review is to compare the effectiveness of single layer GIA (SGIA) versus double layer GIA (DGIA) being used in

general surgery. The particular question we would attempt to answer will be; is single layer hand made GIA in surgical patients is as

effective as double layer?

Search methods

The CCCG (Colorectal Cancer Cochrane Group) Controlled Trials Register, the Cochrane Central Register of Controlled Trials

(CENTRAL) on The Cochrane Library (Issue 1, 2011), MEDLINE (until April 2011) , EMBASE ( The Intelligent Gateway to

Biomedical & Pharmacological Information until April 2011), LILACS (The Latin American and Caribbean Health Sciences Library

until April 2011 ) and Science Citation Index Expanded (SCI-E until April 2011) using the medical subject headings (MeSH) terms

were searched without date, language or age restrictions.

Selection criteria

Randomised, controlled trials comparing the effectiveness of SGIA versus DGIA

Data collection and analysis

At least two review authors independently scrutinised search results, selected eligible studies and extracted data.

Main results

Seven randomised, controlled trials encompassing 842 patients undergoing SGIA versus DGIA were retrieved from the electronic

databases. There were 408 patients in the SGIA group and 432 patients in the DGIA group. All included studies were small, with

sample sizes ranging from 60 to 172. There was no heterogeneity among the included trials. Therefore, in the fixed effects model,

1Single layer versus double layer suture anastomosis of the gastrointestinal tract (Review)


mailto:[email protected]

incidence of anastomotic dehiscence, peri-operative complications and mortality was statistically equivalent between two techniques of

GIA. Average hospital stay following SGIA and DGIA was also comparable. However, SGIA was superior in terms of shorter operative

time. Sensitivity analysis of relatively good quality and poor quality trials supported same conclusion.

Authors’ conclusions

SGIA can be performed quicker as compared to double layer GIA. SGIA is comparable to DGIA in terms of anastomotic leak, peri-

operative complications, mortality and hospital stay. SGIA may routinely be used for GIA following bowel resection. However, since

this conclusion is derived from smaller number of patients recruited in relatively moderate quality trials, further trials should be aimed

to reduce the limitations of this review.

P L A I N L A N G U A G E S U M M A R Y

Single layer versus double layer anastomosis (joining) of the gastrointestinal tract following bowel resection

Bowel anastomosis following resection can be performed in single layer or double layer. This review concludes that single layer

anastomosis is comparable to double layer anastomosis in terms of anastomotic leak, peri-operative complications, death rate and

hospital stay. Single layer anastomosis consumes shorter operative time as compared to double layer. Therefore, single layer anastomosis

may routinely be used for the anastomosis of gastrointestinal tract following bowel resection. However, since this conclusion is derived

from smaller number of patients recruited in relatively moderate quality trials, further trials should be aimed to reduce the limitations

of this review.



SU

MM

AR

YO

FF

IN

DI

NG

SF

OR

TH

EM

AI

NC

OM

PA

RI

SO

N[E

xpla

nati

on]

SinglelayergastrointestinalanastomosiscomparedtoDoublelayergastrointestinalanastomosisforBowelresection

Patientorpopulation:patientswithBowelresection

Settings:

Intervention:Singlelayergastrointestinalanastomosis

Com

parison:Doublelayergastrointestinalanastomosis

Outcomes

Illustrative

comparativerisks*

(95%CI)

Relativeeffect

(95%CI)

NoofParticipants

(studies)

Qualityoftheevidence

(GRADE)

Com

ments

Assum

edrisk

Corresponding

risk

Doublelayergastroin-

testinalanastomosis

Single

layergastroin-

testinalanastomosis

Anastom

oticleak

Oddsratio

Follow-up:

mean

6

months

Studypopulation

OR0.76

(0.44to1.32)

842

(7studies)

⊕©

©©

verylow

1,2

,3

85per1000

66per1000

(39to109)

Low

riskpopulation

Mediumriskpopulation

Anastom

osistime

Meandifference

Follow-up:

mean

6

months

Themeananastomosis

timeinthecontrolgroups

was

0

ThemeanAnastom

osis

timein

theintervention

groupswas

11.12lower

(16.37

to5.37

lower)

218

(2studies)

⊕⊕

©©

low

1,2

,3



http://www.thecochranelibrary.com/view/0/SummaryFindings.html

Lengthofhospitalstay

Meandifference

Follow-up:

mean

6

months

ThemeanLengthofhos-

pitalstay

intheinterven-

tiongroupswas

3.08lower

(8.49

lower

to2.34

higher)

390

(3studies)

⊕⊕

©©

low

4,5

Mortality

Oddsratio

Follow-up:

mean

6

months

Studypopulation

OR0.56

(0.19to1.63)

403

(4studies)

⊕⊕

©©

low

6,7

42per1000

24per1000

(8to67)


Majorcomplications

Oddsratio

Follow-up:

mean

6

months

Studypopulation

OR0.71

(0.44to1.16)

842

(7studies)

⊕⊕

©©

low

8,9

108per1000

79per1000

(51to123)


Anastom

oticleakinhigh

qualitytrials

Oddsratio

Studypopulation

OR1.13

(0.42to3.01)

353

(3studies)

⊕⊕

⊕©

moderate1

0

45per1000

51per1000

(19to124)




Anastom

oticleakinpoor

qualitytrials

Oddsratio

Follow-up:

mean

6

months

Studypopulation

OR0.63

(0.32to1.24)

489

(4studies)

⊕©

©©

verylow

11,1

2,1

3

113per1000

74per1000

(39to136)


*The

basisfortheassumed

risk

(e.g.themediancontrolgroupriskacross

studies)isprovided

infootnotes.Thecorrespondingrisk(and

its95%confidence

interval)isbasedon

the

assumedriskinthecomparison

groupandtherelativeeffectoftheintervention(andits95%CI).

CI:Confidenceinterval;OR:Oddsratio;

GRADEWorkingGroupgradesofevidence

Highquality:Furtherresearchisveryunlikelytochangeourconfidenceintheestimateofeffect.

Moderatequality:Furtherresearchislikelytohaveanimportantimpactonourconfidenceintheestimateofeffectandmaychangetheestimate.

Low

quality:Furtherresearchisverylikelytohaveanimportantimpactonourconfidenceintheestimateofeffectandislikelytochangetheestimate.

Verylowquality:Weareveryuncertainabouttheestimate.

1Thereislackofrandom

isationtechnique,blindness,powercalculationsandintention-to-treatanalysisinsixoutofseventrials

2Prim

aryoutcom

eiscommon

amongalltrialsbutsecondaryoutcom

esvaryconsiderably

3Thereislackofproperrandom

isationtechnique,blinding,allocationconcealment,powercalculations

andintention-to-treatanalysis

4Asabove

5Asabove

6Asabove

7Asabove

8Asabove

9Asabove

10Asabove

11Asabove

12Asabove

13Asabove



B A C K G R O U N D

Description of the condition

In emergency as well as elective situations, gastrointestinal anas-

tomosis (GIA) is an essential step to maintain the continuity of

gastrointestinal tract following intestinal resection secondary to

bowel tumour, intestinal obstruction, blunt/penetrating abdom-

inal trauma, and abdominal sepsis caused by perforated bowel.

GIA is still a source of significant controversy among surgeons

due to the variety of approaches reported for intestinal apposition.

In order to devise a sound and tension free GIA, it is imperative

(Vella 2002) to understand the various clinical, technical and non-

clinical aspects influencing any anastomosis in the gastrointestinal

tract from oesophagus to rectum (Figure 1). Adequate apposition,

appropriate alignment, good local blood supply and tension free

equally spaced stitches can affect GIA positively. Malnutrition,

abdominal sepsis, generalised sepsis, and immunosuppression can

negatively impact on the outcome following GIA (Britton 2003).

The mechanism of GIA healing (Figure 2) is comparable to skin

wound healing which can be divided into lag phase (day 0-3),

fibroplasia phase (day 3-14) and maturation phase (<10 days).

These phases of GIA healing are directly or indirectly influenced

by factors mentioned in Figure 1(McKinley 2006). Anastomotic

leak following GIA has been associated with increased mortality,

morbidity and decreased overall survival in both upper (Griffin

2001; Whooley 2001) and lower gastrointestinal resections (den

Dulk 2009). Anastomotic leak following oesophageal anastomosis

is responsible for 30% to 40% deaths following oesophagectomy

in patients with carcinoma of the thoracic oesophagus (Dai 2009;

Griffin 2001; Whooley 2001). In addition, prolonged hospital

stay, long-term total parenteral nutrition, and intensive care unit

admissions also put massive economic burden on healthcare re-

sources. Since the introduction of total mesorectal excision for rec-

tal cancers, the increased risk of anastomosis leak has been reported

(Carlsen 1998). It is also associated with increased morbidity and

mortality (Hallbook 1996; den Dulk 2009) in the postoperative

period. Rectal anastomotic dehiscence may leads to an increased

local recurrence and poor overall survival in patients undergoing

rectal anastomosis for rectal cancer (Branagan 2005; Law 2007;

Bell 2003; Jung 2008; McArdle 2005; Ptok 2007; Walker 2004).

Therefore, various studies to stratify the risk of GIA dehiscence

and anastomosis techniques to reduce the incidence of leak and

leak-related consequences are important.

Figure 1.



Figure 2.

Historically double layer GIA (DGIA) has been the preferred tech-

nique until the late seventies of the last century. However, dif-

ferent opinions exist about DGIA among surgeons and different

countries in the world. Several published articles highlighted the

risks associated with DGIA (Ceraldi 1993; Deen 1995; Everett

1975; Goligher 1970; Goligher 1977; Irvin 1973; Maurya 1984;

Ordorica 1998; Wayand 1984; Zieren 1993). DGIA is technically

more challenging to perform because it requires the identification

of individual layers of gastrointestinal tract and then approximate

each layer separately making suture tension harder to maintain.

Higher demands of technical skills may result in the increased risk

of errors during the construction of DGIA. Since it takes longer

than single layer GIA (SGIA) construction, so excessive tissue han-

dling can give rise to significant tissue damage and ischemias at the

level of anastomosis resulting in anastomotic dehiscence. Due to

multiple layer closure, it could also reduce the overall intestinal lu-

minal circumference, and thereby prolonging the intestinal recov-

ery following DGIA. The DGIA is more time consuming, because

the surgeon must clear an important segment of each bowel end,

consequently making the procedure more expensive compared to

single layer techniques, not only because it is necessary to use more

suture, but because it takes more time.

As an alternative, SGIA technique took over in early eighties of

the 20th century. SGIA reduces potential risk explained previously

but several authors have suggested that single layer also increases

the risk of dehiscence because the suture technique uses the outer

part of the bowel when it is fashioned with a sero-submucosal



technique (Ballantyne 1984; Burson 1979) or can narrow the in-

testinal lumen when a full thickness technique is used (Azevedo

2005; Singh 1989).

Description of the intervention

GIA is being performed successfully (Burch 2000) for more than

150 years by using various techniques, suture materials, approaches

and devices. This includes double layer inverted technique, dou-

ble layer everted technique, single layer sero-submucal, single layer

full thickness (Irvin 1973; Goligher 1967; Muir 1969; Turnbell

1969) and other innovative and relatively modern techniques of

intestinal anastomosis (Figure 3). Similarly the use of absorbable

versus non-absorbable suture material for GIA has also been advo-

cated in the medical literature (Burch 2000). Monofilament (poly-

dioxanone) and multi filament (silk, polyglactin) suture material

for GIA have been used in few studies without conferring the su-

periority of either suture material (Burch 2000). Since GIA is the

one of the most common procedure being performed in oesoph-

agogastric surgery, hepatobiliary surgery, bariatric surgery, small

bowel surgery and colorectal surgery but it is still unclear whether

single layer or double layer anastomosis are more effective in GIA.

Figure 3.

How the intervention might work

Until now, according to the operating surgeon’s preference both

techniques of GIA are being implied in various surgical specialties

despite the wide spread use of stapling devices. Both techniques

OF GIA has been accepted and being used without knowing the

superiority of either approach. Anastomotic dehiscence is the most

important complication in case of GIA at any level along gas-

trointestinal tract and it should be considered a gold standard to

judge the efficacy of any given anastomosis. Rate of anastomotic

dehiscence following SGIA versus DGIA may guide surgeons that



which technique should be adopted. In addition, other compli-

cations like anastomotic stricture, localized or generalized sepsis

developing due to technical failure of the GIA and time consump-

tion may also help in decision making about adopted approach

for GIA.

Why it is important to do this review

We intend to review the electronic database of the medical lit-

erature and attempt to produce combined and stronger evidence

determining which technique is superior in making the GIA. This

review may help surgeons in future to compare the effectiveness

of both techniques of GIA and adopt the one with lesser risk of

dehiscence and anastomotic complications.

O B J E C T I V E S

The objective of the systematic review is to compare the effective-

ness of SGIA versus DGIA of gastrointestinal tract being used in

various disciplines of surgery. The particular question we would

attempt to answer will be; is single layer hand made gastrointesti-

nal anastomosis in surgical patients of any age and sex using either

absorbable or non-absorbable suture is as effective as double layer

in terms of anastomotic leak, over all morbidity and mortality?

M E T H O D S

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials (RCT’s) comparing the

effectiveness of single layer versus double layer GIA involving oe-

sophageal surgery, gastric surgery, hepatobiliary surgery, bariatric

surgery, small bowel surgery and colorectal surgery. We included

trials on both elective and emergency operations and whatever

the indication was i.e. benign, malignant or traumatic. We in-

cluded trials where anastomosis was either interrupted or continu-

ous. We analysed data from the published randomised trials on all

patients undergoing SGIA versus DGIA as inpatient as well as in

day surgery settings. Patients of any age and gender were included.

Patients of any surgical specialty involving gastrointestinal anasto-

mosis were included in this review. RCT’s (irrespective of type, lan-

guage, blinding, sample size or publication status) that evaluated

the efficacy of SGIA versus DGIA were included. Randomised tri-

als, non-randomised trials, quasi-randomised trials (in which the

methods of allocating participants to a treatment were not strictly

random, such as by date of birth, hospital record number or al-

ternation), retrospective and prospective comparative studies were

reviewed and then decision was made about inclusion or exclu-

sion.

Types of participants

SGIA versus DGIA of the gastrointestinal tract using suture ma-

terial of any type from various surgical disciplines performed by

surgeon of any level of previous experience were reviewed. Out-

comes of GIA given in the randomised trial were compared be-

tween SGIA versus DGIA. We excluded the trials where gastroin-

testinal anastomosis was made using stapling devices or other in-

novative approaches. We attempted to record demographics, rate

of anastomotic leak, rate of stricture formation and anastomosis

related mortality and morbidity in the form of both binary as well

and continuous data.

Types of interventions

We searched the electronic database to find out all published trials

on the effectiveness of SGIA versus DGIA involving oesophageal

surgery, gastric surgery, hepatobiliary surgery, bariatric surgery,

small bowel surgery and colorectal surgery. We extracted data from

the included trials and put it on an Excel spreadsheet. Each trial

quality was assessed and risk of bias was calculated according to

the guidelines given by the Cochrane group for systematic reviews.

Data reported in these trials was analysed to achieve a combined

outcome.

Types of outcome measures

Primary outcomes

Anastomotic leak, diagnosed radiologically or clinically, recorded

as ’number of patients with at least one complication’.

Secondary outcomes

We attempted to analyse all following secondary outcomes if re-

ported in the published randomised, controlled trials.

1-Wound infection: defined according to the Centers for Disease

Control and Prevention (CDC)’s classification (Mangram 1999),

recorded as ’number of complications’, ’types of complications’,

and ’number of patients with at least one complication’.

2-Intra-abdominal infections (peritonitis, abdominal abscess): de-

fined according to the Centers for Disease Control and Prevention

(CDC)’s classification (Mangram 1999), recorded as ’number of

complications’ , ’types of complications’, and ’number of patients

with at least one complication’.

3-Number of re-interventions

4-Overall mortality

5-Overall morbidity

6-Length of hospital stay



7-Operative time

8-Cost analysis

Search methods for identification of studies

Electronic searches

The CCCG (Colorectal Cancer Cochrane Group) Controlled Tri-

als Register, the Cochrane Central Register of Controlled Trials

(CENTRAL) on The Cochrane Library (Issue 1, 2011), MED-

LINE (until April 2011) , EMBASE ( The Intelligent Gate-

way to Biomedical & Pharmacological Information until April

2011), LILACS (The Latin American and Caribbean Health Sci-

ences Library until April 2011 ) and Science Citation Index Ex-

panded (SCI-E until April 2011) using the medical subject head-

ings (MeSH) terms “gastrointestinal anastomosis”, “oesophageal

anastomosis”, “gastric anastomosis”, “small bowel anastomosis”,

“large bowel anastomosis”, “rectal anastomosis”, “single layer gas-

trointestinal anastomosis”, “double layer gastrointestinal anas-

tomosis” and “multiple layer gastrointestinal anastomosis” were

searched. A filter for identifying relevant studies recommended

by the Cochrane Collaboration (Higgins 2008) was used to filter

out irrelevant studies in Medline and EMBASE.

Searching other resources

The references of the included studies were searched to identify

further trials. The ’related article’ function of MEDLINE was also

searched thoroughly in order to identify additional studies. Web-

sites responsible for the registration of the randomised, controlled

trials were searched to find out if there is recent trial running or

ready to publish on this subject. We attempted to gather infor-

mation on all published, unpublished and ongoing trails from all

possible data sources. If necessary, a personal communication by

authors was made to author for correspondence in published trials

for further information on data or clarification. In addition, gas-

trointestinal experts, specialist surgeons and pharmaceutical com-

panies involved in provision of sutures were contacted and asked

to provide details of outstanding clinical trials or any relevant un-

published materials. The international societies of gastrointestinal

surgery were contacted and asked to provide information on any

unpublished studies.

Data collection and analysis

Data was collected on the Excel spread sheet separately by two

authors (MSS, MRSS) and it was further confirmed by the third

author (MKB). The conflict was resolved by mutual agreement

among three authors. We conducted this systematic review accord-

ing to the suggested present protocol and the recommendations by

The Cochrane Reviewers’ Handbook (Higgins 2008). The statis-

tical analysis was performed by MSS and was further confirmed by

MRSS. The software package RevMan 5 (RevMan 5.0.24, 2010)

provided by The Cochrane Collaboration was used for analysis.

The odds ratio (OR) with a 95 percent confidence interval (CI)

was calculated for binary data variables, and the mean difference

(MD) with 95 per cent CI for continuous data variables was cal-

culated. If the mean values were not available for continuous out-

comes, median values were used for the purpose of meta-analysis.

If the standard deviation was not available, it was calculated ac-

cording to the guidelines of The Cochrane Collaboration (Higgins

2008). This involved the assumptions that both groups had the

same variance, which may not be true. The random-effects model

(DerSimonian 1986) and the fixed-effect model (DeMets 1987)

were used to calculate the combined outcome in both binary and

continuous variables. The Mantel-Haenszel method was used for

the calculation of OR under the fixed and random effect models

(Egger 2006). In a sensitivity analysis, 0.5 was added to each cell

frequency for trials in which no event occurred in either the treat-

ment or control group, according to the method recommended by

Deeeks et al (Deeks 2001). The estimate of the difference between

both techniques was pooled, depending upon the affect weights

in results determined by each trial estimate variance. The forest

plot was used for the graphical display of results from the meta-

analysis. The square around the estimate stands for the accuracy

of the estimation (sample size) and the horizontal line represents

the 95% CI. Studies where standard deviation was not reported,

it was estimated either from the range value or P value.

The following details on methods were extracted:

1-Type of anastomosis: Single or double layer, no matter what

layers are used.

2-Type of suture used. Absorbable, non absorbable, monofilament

or multi filament

3-Intestinal segment: oesophagus, stomach, duodenum, jejunum,

ileum, colon and rectum

The following data on randomisation and blinding procedure will

be extracted:

1-Number of randomised patients.

2-Number of patients not randomised and reasons for non-ran-

domisation.

3-Exclusion after randomisation.

4-Drop-outs.

5-Blinding of patients and observers.

6-’Intention-to-treat’ analysis.

7-Internal validity.

Selection of studies

Studies were selected according to the predefined inclusion crite-

ria. We analysed all published randomised, controlled trial report-

ing the effectiveness of SGIA versus DGIA in the gastrointestinal

tract from oesophagus to rectum in terms of luminal continuity,



anastomotic leak, stenosis and morbidity.

Data extraction and management

Data was collected on the Excel spread sheet separately by two

authors (MSS, MRSS) and it was further confirmed/re-checked

by the third author (MKB). The conflict regarding data mismatch

was resolved by mutual agreement among three authors.

Assessment of risk of bias in included studies

We defined the methodological quality as the confidence that

the design and report restrict bias in the intervention compari-

son (Chalmers 1998; Higgins 2008; Jadad 1996; Moher 1998;

Kjaergard 2001) and four important parameters (randomisation

technique, allocation concealment, blinding, intention to treat

analysis) for a high quality randomised controlled trial were clearly

described in the reported study. We also looked for power calcu-

lation and strength of the trial in order to score it precisely and

accurately. Due to the risk of overestimation of intervention ef-

fects in randomised trials with inadequate methodological quality

(Chalmers 1998; (Higgins 2008; Jadad 1996; Schulz 1995; Moher

1998; Kjaergard 2001), we assessed the influence of methodolog-

ical quality as follows.

Generation of the allocation sequence. Adequate, if the allocation

sequence was generated by a computer or random number table.

Drawing of lots, tossing of a coin, shuffling of cards, or throwing

dice was considered as adequate if a person who was not other-

wise involved in the recruitment of participants performed the

procedure. Unclear, if the trial was described as randomised, but

the method used for the allocation sequence generation was not

described. Inadequate, if a system involving dates, names, or ad-

mittance numbers were used for the allocation of patients. These

studies were known as quasi-randomised and was excluded from

the present review when assessing beneficial effects.

Allocation concealment. Adequate, if the allocation of patients

involved a central independent unit, on-site locked computer, or

sealed envelopes. Unclear, if the trial was described as randomised,

but the method used to conceal the allocation was not described.

Inadequate, if the allocation sequence was known to the inves-

tigators who assigned participants or if the study was quasi-ran-

domised.

Double/ blinding or masking. Adequate, if the trial was described

as double blind and the method of blinding was described. Un-

clear, if the trial was described as double blind, but the method

of blinding was not described. Not performed, if there was no

blinding at all.

Follow-up/intention to treat analysis. Adequate, if the numbers

and reasons for dropouts and withdrawals in all intervention

groups were described or if it was specified that there were no

dropouts or withdrawals. Unclear, if the report gave the impres-

sion that there had been no dropouts or withdrawals, but this was

not specifically stated. Inadequate, if the number or reasons for

dropouts and withdrawals were not described.

Measures of treatment effect

We scored each study according to previously recommended tech-

niques (Chalmers 1998; Higgins 2008; Jadad 1996) in order to

define risk of bias, power of the study, presence or absence of blind-

ing and calculations based on the intention to treat analysis. Each

trial was scored as follows: score A (adequate), score B (unclear),

score C (not concealed) and score D (partially not done). Trials

scoring A and B were preferably included and trials scoring C and

D were included when further good quality studies could not be

retrieved.

Unit of analysis issues

The RR with a 95 percent CI was calculated for binary data vari-

ables, and the mean difference (MD) with a 95 per cent CI for

continuous data variables was calculated. If the mean values were

not available for continuous outcomes, median values were used

for the purpose of meta-analysis. If the standard deviation was

not available, it was calculated according to the guidelines of the

Cochrane Collaboration (Higgins 2008). This involved the as-

sumptions that both groups had the same variance, which may

not be true. The random-effects model (DerSimonian 1986) and

the fixed-effect model (DeMets 1987) were used to calculate the

combined outcome in both binary and continuous variables. The

Mantel-Haenszel method was used for the calculation of RR under

the fixed and random effect models (Egger 2006). In a sensitivity

analysis, 0.5 was added to each cell frequency for trials in which no

event occurred in either the treatment or control group, according

to the method recommended by Deeks et al (Deeks 2001). The

estimate of the difference between both techniques was pooled,

depending upon the affect weights in results determined by each

trial estimate variance. The forest plot was used for the graphical

display of results from the meta-analysis. The square around the

estimate stands for the accuracy of the estimation (sample size)

and the horizontal line represents the 95% CI.

Dealing with missing data

We contacted the first author via personal communication in order

to get missing data. When further information were required from

any source we contacted every relevant person involved in running

of the published trial. When missing data could not be achieved

and that particular trial did not score according to our inclusion

criteria we excluded that trial and explained it in the table giving

details about excluded trials.



Assessment of heterogeneity

In case of heterogeneity, only the results of the random-effects

model were reported. Heterogeneity was explored using the chi2 test, with significance set at P < 0.05, and it was quantified

(Higgins 2002) using I2, with a maximum value of 30 percent

identifying low heterogeneity (Higgins 2008).

Assessment of reporting biases

We defined the methodological quality as the confidence that

the design and report restrict bias in the intervention compari-

son (Chalmers 1998; Higgins 2008; Jadad 1996; Moher 1998;

Kjaergard 2001) and four important parameters (randomisation

technique, allocation concealment, blinding, intention to treat

analysis) for a high quality randomised controlled trial were clearly

described in the reported study. We also looked for power calcu-

lation and strength of the trial in order to score it precisely and

accurately. Due to the risk of overestimation of intervention effects

in the randomised trials with inadequate methodological quality

(Chalmers 1998; Higgins 2008; Jadad 1996; Schulz 1995; Moher

1998; Kjaergard 2001), we assessed the influence of methodolog-

ical quality.

Data synthesis

Data of all primary and secondary outcomes was synthesized on

the Revman 5.1.2 provided by the Cochrane Collaboration in

order to achieve a summative outcome.

Subgroup analysis and investigation of heterogeneity

We attempted to study if some difference exists between absorbable

or non-absorbable sutures in making the GIA measuring their ef-

fectiveness in terms of primary and secondary outcomes. In addi-

tion we also attempted to study different surgical specialties sep-

arately involving gastrointestinal anastomosis. We attempted to

perform a sensitivity analysis in order to compare the intervention

effect in trials with high methodological quality (i.e. trials with

adequate generation of the allocation sequence, allocation con-

cealment, and blinding) to that of trials with low methodological

quality (i.e., trials not having one or more adequate component).

We also attempted to perform subgroup analyses according to the

gastrointestinal segment included in the trials, type of suture used,

as well as type of procedure (elective versus acute). Further, we

explored causes of heterogeneity (defined as the presence of statis-

tical heterogeneity by chi-squared test with significance set at P-

value less than 0.10 and measure the quantities of heterogeneity

by I2 (Higgins 2002) and by comparing different groups of trials

stratified according to patient risk factors, level of experience of the

surgeon, and other factors that may explain heterogeneity. Clinical

and methodological causes of the heterogeneity were searched in

reported trials and it was be clearly documented in this review.

Sensitivity analysis

Sensitivity analysis was attempted by using funnel plot in order

to determine potential bias in the reported trial and isolate the

outliers.

R E S U L T S

Description of studies

See: Characteristics of included studies; Characteristics of excluded

studies.

See:Characteristics of included studies and Characteristics of ex-

cluded studies

Results of the search

We searched 6 electronic databases using MeSH headings men-

tioned in methods section and search strategies explained in Ap-

pendices 1-4. The search of CENTRAL database produced 51

relevant trials. Similarly search of MEDLINE, EMBASE, LILI-

ACS, LILACS and SCI-E produced 53, 44, 29, and 17 relevant

trials potentially suitable for inclusion for meta-analysis of SGIA

versus DGIA techniques. There were combined 59 studies mu-

tually retrieved for the electronic databases. Further screening of

these potentially relevant studies was performed comprehensively

as explained in Figure 4. Seven studies (Burch 2000; Everett 1975;

Goligher 1977; Irvin 1973; Maurya 1984; Ordorica-Flores 1998;

Wayand 1984) encompassing 842 patients undergoing SGIA ver-

sus DGIA were found suitable for inclusion in this review.



Figure 4. Quorum diagram showing trial selection methodology

Included studies

Seven randomised, controlled trials (Burch 2000; Everett 1975;



sus DGIA were retrieved from the above mentioned 6 electronic

databases. There were 408 patients in the SGIA group and 432

patients in the DGIA group. Recruited patients in included trials

had not had clinical, biochemical and radiological evidence of en-

terocutaneous or colocutaneous fistula preoperatively.All included

studies were small, with sample sizes ranging from 60 to 172. All

participants were considered to have gastrointestinal pathology re-

quiring either SGIA or DGIA in order to maintain gastrointestinal

continuity following surgery. The recruitment group underwent

gastric anastomosis, small bowel to small bowel anastomosis, small

bowel to colon anastomosis, colon to colon anastomosis and colon

to rectum anastomosis. It also included the group of patients un-

dergoing reversal of ileostomy and colostomy. Individual number

of type and level of the anastomoses in both limbs of the ran-

domised trials were not reported. Pre-operative risk stratification

for anastomotic dehiscence was not reported in all studies. None

of the reported trial was multicenter.

Burch et al (Burch 2000) recruited 132 patients reporting 65 pa-

tients undergoing SGIA with 3/0 polypropylene and 67 patients

undergoing DGIA with 3/0 silk and 3/0 polyglycolic acid. Re-

cruited patients underwent enteroenterostomy, enterocolostomy

and colostomy. The preoperative confounding interventions in the

both limbs of the trial were prophylactic antibiotics, dietary restric-

tions, mechanical bowel preparation and on table rectal wash. This

trial was run in the USA. Primary outcome was anastomotic leak

and secondary outcomes included intra-abdominal abscess forma-

tion, operative time, peri-operative complications and length of

hospital stay.

Everett (Everett 1975) recruited 92 patients reporting 40 patients

undergoing SGIA with 4/0 supramid and 52 patients undergo-

ing DGIA with 2/0 chromic catgut and 4/0 supramid. Recruited

patients underwent upper and lower rectal anastomosis. The pre-

operative confounding interventions in the both limbs of the

trial were prophylactic antibiotics, dietary restrictions, mechanical

bowel preparation and on table rectal wash. This trial was run in



the UK. Primary outcome was anastomotic leak and secondary

outcomes included were peri-operative complications and mortal-

ity.

Goligher et al Goligher 1977) recruited 133 patients reporting

69 patients undergoing SGIA with 4/0 supramid and 66 patients

undergoing DGIA with 2/0 chromic catgut and 3/0 silk. Re-

cruited patients underwent upper and lower rectal anastomosis.

The preoperative confounding interventions in the both limbs

of the trial were prophylactic antibiotics, dietary restrictions, me-

chanical bowel preparation and on table rectal wash. This trial

was run in the UK. Primary outcome was anastomotic leak and

secondary outcome was peri-operative complications.

Irvin et al (Irvin 1973) recruited 60 patients reporting 29 patients

undergoing SGIA with 3/0 silk and 31 patients undergoing DGIA

with 3/0 chromic catgut and 3/0 silk. Recruited patients under-

went ileo-ileal, ileo-colic, colo-colic and colo-rectal anastomosis.

The preoperative confounding interventions in the both limbs

of the trial were prophylactic antibiotics, dietary restrictions, me-

chanical bowel preparation and on table rectal wash. This trial

was run in the UK. Primary outcome was anastomotic leak and

secondary outcomes included were peri-operative complications

and mortality.

Maurya et al (Maurya 1984) recruited 172 patients reporting 60

patients undergoing SGIA with interrupted 3/0 silk and 112 pa-

tients undergoing DGIA with 3/0 chromic catgut and 3/0 silk.

Recruited patients underwent ileo-ileal, ileo-colic, and colo-colic

anastomosis. The preoperative confounding interventions were

not reported, however, all recruited patients were acute surgical

admissions. This trial was run in the India. Primary outcome was

anastomotic leak and secondary outcomes included were peri-op-

erative complications and hospital stay.

Ordorica-Flores et al (Ordorica-Flores 1998) recruited 86 patients

reporting 42 patients undergoing SGIA with 4/0 or 5/0 polyglactin

silk and 44 patients undergoing DGIA with 4/0 or 5/0 polyglactin.

Recruited patients underwent ileo-ileal, ileo-colic, colo-colic, colo-

rectal anastomosis, ileostomy closure and colostomy closure. The

preoperative confounding interventions in the both limbs of the

trial were prophylactic antibiotics, dietary restrictions, mechanical

bowel preparation and on table rectal wash. This trial was run in

the Mexico. Primary outcome was anastomotic leak and secondary

outcomes included were peri-operative complications, operative

time, hospital stay and mortality.

Wayand et al (Wayand 1984) recruited 165 patients reporting

103 patients undergoing SGIA with 3/0 or 4/0 polyglactin silk

and 62 patients undergoing DGIA with 3/0 polyglactin and 2/

0 catgut. Recruited patients underwent ileo-ileal, ileo-colic, colo-

colic, colo-rectal anastomosis, and gastric anastomosis. The pre-

operative confounding interventions were not reported. This trial

was run in the Germany. Primary outcome was anastomotic leak

and secondary outcomes included were peri-operative complica-

tions and mortality.

Excluded studies

There were two (Carty 1991; Kingsnorth 1989) excluded studies

which were prospective trials but without randomisations and any

trial protocol.

Risk of bias in included studies

Also see:Characteristics of included studies

Allocation

Allocation of the recruited patients in included trials was not re-

ported precisely. For a relatively good quality randomised, con-

trolled trial optimum sequence generation and relatively adequate

randomisation was reported in three (Burch 2000; Goligher 1977;

Ordorica-Flores 1998) trials only where participants were ran-

domly distributed to the control or experimental group accord-

ing to a computer-generated code. We classified high risk of bias

due to poor randomisations technique and inadequate allocation

concealment in case of Goligher et al (Goligher 1977), Irvin et al

Irvin 1973), and Maurya et al (Maurya 1984). Unclear risk of

bias was classified in case of one trial (Wayand 1984) due to the

absence of randomisation technique reporting and lack of alloca-

tion concealment.

Blinding

Blinding of the operating surgeon was not possible due to nature

of the trial. However, blinding of the trail participants and out-

come assessor was neither investigated nor reported by any of the

included study except one trial (Ordorica-Flores 1998). Therefore,

based on blinding technique six included studies were classified

inadequate and high risk.

Incomplete outcome data

In judging the risk of bias for incomplete data reporting, we eval-

uated primary outcome measure; namely, anastomotic leak. We

also considered whether an intention-to treat (ITT) analysis was

considered the primary outcome and whether missing data were

imputed appropriately. We did not find risk of bias due to incom-

plete outcome data. Therefore, all included studies were classified

moderate to low risk for incomplete data reporting. Lack of ITT

was not reported in all trials

Selective reporting

In judging the risk of bias for selective reporting, we were unable

to assess the trial protocols and therefore assessed the studies based

on the pre-specified outcome measures reported in the methods

section of the trial report. The risk of bias due to selective reporting

was considered low for all seven trials as all of the pre-specified

outcomes were reported.



Other potential sources of bias

There have been published studies suggesting that industry-spon-

sored trials may overestimate the treatment effect (Bhandari 2004;

Thomas 2008). None of the included trial reported any sponsor-

ing body. Four trials (Everett 1975; Goligher 1977; Irvin 1973;

Ordorica-Flores 1998) reported acknowledgement to statistician

and assisting nursing staff. The abstract of one study (Burch 2000)

was presented at Annual Meeting of the Professional Society be-

fore publication. Approval of the study by local/regional Research

Ethics Committee was not reported adequately by all trials.

Effects of interventions

See: Summary of findings for the main comparison Single

layer gastrointestinal anastomosis compared to Double layer

gastrointestinal anastomosis for Bowel resection

PRIMARY OUTCOME

Anastomotic leak

There was no heterogeneity (chi2 = 4.13, df = 6, (P = 0.66); I2 =

0%) among included trials. Therefore, in the fixed effects model

(OR, 0.76; 95% CI, 0.44, 1.32; z = 0.98; P = 0.33; Figure 5), the

risk of anastomotic dehiscence following SGIA was slightly less

than DGIA but statistically it was not significant. Both techniques

were equally effective for GIA.

Figure 5. Forest plot of comparison: 1 Anastomosis, outcome: 1.1 Anastomosis leak.

SECONDARY OUTCOMES

Peri-operative complications

There was no heterogeneity (chi2 = 3.65, df = 6, (P = 0.72); I2 =

0%) among included trials. Therefore, in the fixed effects model

(OR, 0.76; 95% CI, 0.44, 1.32; z = 1.37; P = 0.17; Figure 6), the

risk of peri-operative complications following SGIA and DGIA

was statistically same.

Figure 6. Forest plot of comparison: 1 Anastomosis, outcome: 1.2 Anastomosis time.



Mortality

Four trials (Everett 1975; Irvin 1973; Ordorica-Flores 1998;

Wayand 1984) contributed in the combined outcome of overall

mortality following GIA. There was no heterogeneity (chi2 = 0.17,

df = 3, P = 0.98; I2 = 0%) among included trials. Therefore, in the

fixed effects model (OR, 0.56; 95% CI, 0.19, 1.63; z = 1.07; P =

0.29; Figure 7), the risk of mortality following SGIA and DGIA

was statistically same.

Figure 7. Forest plot of comparison: 1 Anastomosis, outcome: 1.3 Hospital stay.

Hospital stay

Three trials (Burch 2000; Maurya 1984; Ordorica-Flores 1998)

contributed in the combined outcome of hospital stay following

GIA. There was significant heterogeneity (Tau2 = 21.82, chi2 =

100.57, df = 2, P < 0.00001; I2 = 98%) among included three

trials. Therefore, in the random effects model (MD, -3.08; 95%

CI, -8.49, 2.34; z = 1.11; P = 0.27; Figure 8), the average length

of hospital stay following SGIA and DGIA was statistically same.

Figure 8. Forest plot of comparison: 1 Anastomosis, outcome: 1.4 Mortality.



Operative time of the anastomosis

Two trials (Burch 2000; Ordorica-Flores 1998) contributed in the

combined outcome of operative time. There was no heterogeneity

(chi2 = 1.0, df = 1, (P = 0.32); I2 = 0%) between two included

trials. Therefore, in the fixed effects model (MD, -11.2; 95% CI,

-16.37, -5.87; z = 4.15; P = 0.00001; Figure 9), average operative

time for SGIA was significantly shorter than the DGIA.Figure 9. Forest plot of comparison: 1 Anastomosis, outcome: 1.5 Major complications.

SENSITIVITY ANALYSIS

Anatomotic leak among high quality trials

Three trials ( Burch 2000; Goligher 1977; Ordorica-Flores 1998)

were classified as high quality studies. There was no heterogeneity

(chi2 = 0.55, df = 2, (P = 0.76); I2 = 0%) among included trials.

Therefore, in the fixed effects model (OR, 1.13; 95% CI, 0.42,

3.01; z = 0.24; P = 0.81; Figure 10), the risk of anastomotic de-

hiscence following SGIA was as high as in DGIA.

Figure 10. Forest plot of comparison: 1 Anastomosis, outcome: 1.6 Anatomotic leak in high quality trials.



Anatomotic leak among poor quality trials

Four trials (Everett 1975; Irvin 1973; Maurya 1984; Wayand

1984) were classified as poor quality studies.There was no hetero-

geneity (chi2 = 2.90, df = 3, (P = 0.41); I2 = 0%) among included

trials. Therefore, in the fixed effects model (OR, 0.63; 95% CI,

0.32, 1.24; z = 1.33; P = 0.18; Figure 11), the risk of anastomotic

dehiscence following SGIA was as high as in the DGIA.

Figure 11. Forest plot of comparison: 1 Anastomosis, outcome: 1.7 Anastomotic leak in poor quality trials.

D I S C U S S I O N

Summary of main results

Seven randomised, controlled trials (Burch 2000; Everett 1975;



sus DGIA were retrieved from the electronic databases. There were

408 patients in the SGIA group and 432 patients in the DGIA

group. Recruited patients in included trials had not had clinical,

biochemical and radiological evidence of enterocutaneous or colo-

cutaneous fistula preoperatively. All included studies were small,

with sample sizes ranging from 60 to 172. Pre-operative risk strati-

fication for anastomotic dehiscence was not reported in all studies.

None of the reported trial was multicenter. There was no hetero-

geneity among included studies. All of the included studies inves-

tigated anastomotic leak as a primary outcome. In the fixed ef-

fects model, incidence of anastomotic dehiscence following SGIA

was slightly less than DGIA but statistically it was not significant.

Among secondary outcomes there was no difference in the inci-

dence of mortality, peri-operative complications and length of hos-

pital stay between both techniques of GIA. However, as expected

SGIA takes shorter time to construct as compared to DGIA. Sub-

group and sensitivity analysis of relatively good quality and poor

quality trials separately supported same conclusion.

Overall completeness and applicability ofevidence

All trials evaluated primary outcome of anastomotic leak according

to pre-trial analysis strategy. The utilisation of anastomotic leak as

a primary endpoint following SGIA and DGIA was well targeted

because leak occupies the greatest attention among surgical frater-

nity. surgeons. Primary outcome was thoroughly investigated and

adequately reported. Summated outcome of the primary variable

is conclusive and may be considered adequate. However, due to

inadequate quality of randomised, controlled trials, the applica-

tion of this conclusion in current clinical settings for GIA does

not carry heavy weight. Routine use of SGIA may be considered

until a high quality trial is available to agree or disagree with our

conclusion.

Quality of the evidence

There is lack of adequate randomisation technique, allocation con-

cealment, single or double blinding (except one trial; Ordorica-

Flores 1998),ITT and power of the study in included trials. Based

on this, authors consider the quality of the evidence resulting from

this review is inadequate.

Potential biases in the review process

There are several limitations in this review. Firstly, the study by

Maurya et al (Maurya 1984) had substantial influence on the com-



bined OR and effect weight of the meta-analysis (44.1%). Con-

sidering it is a poor quality randomised trial from a relatively un-

known centre and published in low impact journal, it is a strong

source of contamination and bias in this review. Additionally, over

all effect weight of poor quality (Everett 1975; Irvin 1973; Maurya

1984; Wayand 1984) trials is 75 % leaving the effect weight of

relatively good quality trials around 25%. Therefore, this conclu-

sion is mainly based on the summative outcome of poor quality

trials. Secondly, the quality of included trials was not necessarily

high due to the lack of adequate randomisation technique, allo-

cation concealment, single or double blinding (except one trial;

Ordorica-Flores 1998), ITT and power of the study which are po-

tential sources of higher degree of bias. Thirdly, there have been sig-

nificant differences about inclusion and exclusion criteria among

included trials. Fourthly, variable degree of differences also existed

among included trials about the definition of “anastomotic leak”.

One study (Goligher 1977) included subclinical leak in their data

analysis and reporting while majority of the trials reported clinical

leak requiring further surgical or radiological intervention. Fifthly,

fewer studies recruiting a very small number of patients in this re-

view may not be sufficient to recognise small differences between

SGIA and DGIA. Lastly, because there was no difference in pri-

mary outcomes between two techniques, choices in trials should

have been made after taking into account the results of other out-

comes such as overall mortality, duration of operative procedure,

length of hospital stay, wound infection rate and cost analysis.

Agreements and disagreements with otherstudies or reviews

Shikata et al (Shikata 2006) has published a meta-analysis of six

randomised, controlled trials. The results of that meta-analysis

support our conclusion conferring DGIA offers no definitive ad-

vantage over SGIA as for as anastomotic leak is concerned. They

also recommended the routine use of SGIA considering shorter

duration of procedure and being economical, despite not report-

ing the results on cost effectiveness of either technique. They did

not report data on the summated conclusion of secondary out-

comes. Among included trials, 4 studies (Burch 2000; Irvin 1973;

Maurya 1984; Ordorica-Flores 1998) reported slightly higher in-

cidence of anastomotic leak rate following SGIA but statistically

it was not significant.

A U T H O R S ’ C O N C L U S I O N S

Implications for practice

At present, there is no high quality evidence available to suggest

that SGIA is superior to DGIA in order to maintain the continuity

of gastrointestinal tract following bowel resection. The conclusion

of this review is based on 3 relatively good quality trials and 4 poor

quality trials. Despite having several limitations in this review,

we still believe that our meta-analysis provides the current best

available evidence in making the clinical decision to choose either

SGIA or DGIA. SGIA may routinely be used for GIA following

bowel resection until a stronger evidence is reported.

Implications for research

The conclusion of this review opens the channels for further re-

search on this very important area of gastrointestinal surgery. A

major multicentre randomised controlled trial of high quality is

required in order to strengthen current evidence. Trials on high

risk patients (rectal anastomosis versus ileal anastomosis) should

be performed to find which group may benefit more from which

type of anastomosis. Studies on the use of SGIA versus DGIA in

patients undergoing anastomosis of different levels of gastrointesti-

nal tract should also be performed separately in order to exclude

the effect of confounding and adjunctive procedures. Further re-

search on the effect of combined use of recently introduced and

improved open as well as laparoscopic stapling devices along with

suture anastomosis of gastrointestinal tract needs an extensive ex-

ploration. Methodologically sound and robust randomised, con-

trolled trials are needed in order to investigate further any effect of

SGIA or DGIA using various types of sutures like absorbable ver-

sus non-absorbable, mono-filamentous versus multi-filamentous,

and continuous versus interrupted stitches. When conducting and

reporting randomised, controlled trials the investigators should

follow the CONSORT statement for reporting controlled trials

(CONSORT 2010) so that the randomised, controlled trials can

be precisely and accurately evaluated by readers and reviewers.

A C K N O W L E D G E M E N T S

We are very grateful to Ms Ann Alidina Specialist Colorectal Nurse

for providing us valuable information about published trials and

verifying the extracted data as an external reviewer.



R E F E R E N C E S

References to studies included in this review

Burch 2000 {published data only}

Burch JM, Franciose RJ, Moore EE, Biffl WL, Offner

PJ. Single-layer continuous versus two-layer interrupted

intestinal anastomosis: a prospective randomized trial. Ann

Surg 2000;231:832–37.

Everett 1975 {published data only}

Everett WG. A comparison of one layer and two layer

techniques for colorectal anastomosis. Br J Surg 1975;62:

135–40.

Goligher 1977 {published data only}

Goligher JC, Lee PW, Simpkins KC, Lintott DJ. A

controlled comparison one- and two-layer techniques of

suture for high and low colorectal anastomoses. Br J Surg

1977;64:609–14.

Irvin 1973 {published data only}

Irvin TT, Goligher JC, Johnston D. A randomized

prospective clinical trial of single-layer and two-layer

inverting intestinal anastomoses. Br J Surg 1973;60:

457–60.

Maurya 1984 {published data only}

Maurya SD, Gupta HC, Tewari A, Khan SS, Sharma BD.

Double layer versus single layer intestinal anastomosis: a

clinical trial. Int Surg 1984;69:339–40.

Ordorica-Flores 1998 {published data only}

Ordorica-Flores RM, Bracho-Blanchet E, Nieto-Zermeño

J, Reyes-Retana R, Tovilla-Mercado JM, Leon-Villanueva

V, Varela-Fascinetto G. Intestinal anastomosis in children:

a comparative study between two different techniques. J

Pediatr Surg 1998;33:1757–59.

Wayand 1984 {published data only}

Wayand W, Rieger R, Umlauft M. Single or double layer? A

controlled prospective study on the comparison of 2 suture

technics in gastrointestinal anastomoses. Chirurg 1984;55:

650–52.

References to studies excluded from this review

Carty 1991 {published data only}

Carty NJ, Keating J, Campbell J, Karanjia N, Heald RJ.

Prospective audit of an extramucosal technique for intestinal

anastomosis. Br J Surg 1991;78:1438–41.

Kingsnorth 1989 {published data only}

Kingsnorth AN, Makin CA, Ellenbogen S. Prospective study

of the serosubmucosal (extramucosal) suture technique for

gastrointestinal anastomosis. J R Coll Surg Edinb 1989;34:

130–32.

Additional references

Azevedo 2005

Azevedo JL, Da Silva CE, Azevedo OC, Simoes Mde J. One

layer sutures of digestive tract knotted in the lumen, in dogs:

perforating stitch versus serosubmucosal suture. Acta Cir

Bras 2005;20(2):168–173.

Ballantyne 1984

Ballantyne GH. The experimental basis of intestinal

suturing. Effect of surgical technique, inflammation, and

infection on enteric wound healing. Dis Colon Rectum

1984;27(1):61–71.

Bell 2003

Bell SW, Walker KG, Rickard MJ, Sinclair G, Dent OF,

Chapuis PH, Bokey EL. Anastomotic leakage after curative

anterior resection results in a higher prevalence of local

recurrence. Br J Surg 2003;90(10):1261–1266.

Bhandari 2004

Bhandari M, Busse JW, Jackowski D, Montori VM,

SchünemannH, Sprague S. Association between industry

funding and statistically significant pro-industry findings in

medical and surgical randomized trials. Canadian Medical

Association Journal 2004;170:477–480.

Branagan 2005

Branagan G, Finnis D. Wessex Colorectal Cancer Audit

Working Group. Prognosis after anastomotic leakage

in colorectal surgery. Dis Colon Rectum 2005;48(5):

1021–1026.

Britton 2003

Britton J. Intestinal Anastomosis. In: Julian Britton editor

(s). ACS Surgery: Principles and Practice; 5-Gastrointestinal

Tract and Abdomen; 29 Intestinal Anastomosis. 1. WebMD,

2003:29.

Burson 1979

Burson LC, Berliner SD, Strauss RJ, Katz P, Wise L.

Telescoping anastomosis of the colon: a comparative study.

Dis Colon Rectum 1979;22(2):111–116.

Carlsen 1998

Carlsen E, Schlichting E, Guldvog I, Johnson E, Heald RJ.

Effect of the introduction of total mesorectal excision for the

treatment of rectal cancer.. Br J Surg 1998;85(4):526–529.

Ceraldi 1993

Ceraldi CM, Rypins EB, Monahan M, Chang B, Sarfeh

IJ. Comparison of continuous single layer polypropylene

anastomosis with double layer and stapled anastomoses in

elective colon resections. Am Surg 1993;59(3):168–171.

Chalmers 1998

Chalmers TC, Smith H, Jr, Blackburn B, Silverman B,

Schroeder B, Reitman D, Ambroz A. A method for assessing

the quality of a randomized control trial. Control Clin Trials

1998;2:31–49.

CONSORT 2010

Moher D, Hopewell S, Schulz KF, Montori V, Gøtzsche

PC, Devereau PJ, et al.CONSORT 2010 explanation and

elaboration:updated guidelines for reporting parallel group

randomised trials. BMJ 2010 Mar 23;340:DOI: 10.1136/

BMJ:c869.

Dai 2009

Dai YY, Gretschel S, Dudeck O, Rau B, Schlag PM,

Hunerbein M. Treatment of oesophageal anastomotic leaks



by temporary stenting with self-expanding plastic stents. Br

J Surg 2009;98(8):887–891.

Deeks 2001

Deeks JJ, Altman DG, Bradburn MJ. Statistical methods for

examining heterogeneity and combining results from several

studies in meta-analysis. Systemic reviews in health care: meta-

analysis in context. 2nd Edition. London: BMJ Publication

Group, 2001.

Deen 1995

Deen KI, Smart PJ. Prospective evaluation of sutured,

continuous, and interrupted single layer colonic

anastomoses. Eur J Surg 1995;161(10):751–753.

DeMets 1987

DeMets DL. Methods for combining randomized clinical

trials: strengths and limitations. Stat Med 1987;6(3):

341–350.

den Dulk 2009

den Dulk M, Marijnen CA, Collette L, Putter H, Påhlman

L, Folkesson J, Bosset JF, Rödel C, Bujko K, van de

Velde CJ. Multicentre analysis of oncological and survival

outcomes following anastomotic leakage after rectal cancer

surgery.. Br J Surg 2009;96(9):1066–1075.

DerSimonian 1986

DerSimonian R, Laird N. Meta-analysis in clinical trials.

Control Clin Trials 1986;7(3):177–188.

Egger 2006

Egger M, Smith GD, Altman DG. Systematic reviews in

healthcare. London: BMJ Publication Group, 2006.

Goligher 1967

Goligher JC. Surgery of the Anus, Rectum and Colon. In:

Goligher JC editor(s). Surgery of the Anus, Rectum and

Colon. 1. London: Bailliere, Tindall & Cassell, 1967:522.

Goligher 1970

Goligher JC, Morris C, McAdam WA, De Dombal FT,

Johnston D. A controlled trial of inverting versus everting

intestinal suture in clinical large-bowel surgery. Br J Surg

1970;57(11):817–822.

Griffin 2001

Griffin SM, Lamb PJ, Dresner SM, Richardson DL, Hayes

N. Diagnosis and management of a mediastinal leak

following radical oesophagectomy. Br J Surg 2001;88(10):

1346–1351.

Hallbook 1996

Hallbook O, Sjodahl R. Anastomotic leakage and functional

outcome after anterior resection of the rectum. Br J Surg

1996;83(1):60–62.

Higgins 2002

Higgins JP, Thompson SG. Quantifying heterogeneity in a

meta-analysis. Stat Med 2002;21(11):1539–1558.

Higgins 2008

Higgins JPT, Green S (editors). Cochrane Handbook for

Systematic Reviews of Interventions Version 5.0.0. Vol. http:

//www.cochrane-handbook.org [accessed on 4th July

2010], York: Cochrane Collaboration, 2008.

Jadad 1996

Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds

DJ, Gavaghan DJ, McQuay HJ. Assessing the quality of

reports of randomized clinical trials: is blinding necessary?.

Control Clin Trials 1996;17:1–12.

Jung 2008

SH, Yu CS, Choi PW, Kim DD, Park IJ, Kim HC, Kim JC.

Risk factors and oncologic impact of anastomotic leakage

after rectal cancer surgery. Dis Colon Rectum 2008;51(6):

902–908.

Kjaergard 2001

Kjaergard LL, Villumsen J, Gluud C. Reported

methodologic quality and discrepancies between large and

small randomized trials in meta-analyses. Ann Intern Med

2001;135(11):982–989.

Law 2007

Law WL, Choi HK, Lee YM, Ho JW, Seto CL. Anastomotic

leakage is associated with poor long-term outcome in

patients after curative colorectal resection for malignancy. J

Gastrointest Surg 2007;11(1):8–15.

Mangram 1999

Mangram AJ, Horan TC, Pearson ML, Silver LC, Jarvis

WR. The Hospital Infection Control Practices Advisory

Committee. Guideline for the prevention of surgical site

infection. Centers for Disease Control and Prevention,

1999.

McArdle 2005

McArdle CS, McMillan DC, Hole DJ. Impact of

anastomotic leakage on long-term survival of patients

undergoing curative resection for colorectal cancer. Br J

Surg 2005;92(9):1150–1154.

McKinley 2006

McKinley AJ, Kukowski ZH. Intestinal anastomosis.

Surgery 2006;24(7):224–229.

Moher 1998

Moher D, Pham B, Jones A, Cook DJ, Jadad AR, Moher M,

Tugwell P, Klassen TP. Does quality of reports of randomised

trials affect estimates of intervention efficacy reported in

meta-analyses?. Lancet 1998;352(9128):609–613.

Muir 1969

Muir EG. Operative Surgery. In: Muir EG editor(s).

Operative Surgery. 1. Vol. 5, London: BMJ publishers,

1969:655.

Ordorica 1998

Ordorica-Flores RM, Bracho-Blanchet E, Nieto-Zermeno J,

Reyes-Retana R, Tovilla-Mercado JM, Leon-Villanueva V,

Varela-Fascinetto G. Intestinal anastomosis in children: a

comparative study between two different techniques. J Pead

Surg 1998;33(12):1757–1759.

Ptok 2007

Ptok H, Marusch F, Meyer F, Schubert D, Gastinger I,

Lippert H. Impact of anastomotic leakage on oncological

outcome after rectal cancer resection. Br J Surg 2007;94

(12):1548–1554.



Schulz 1995

Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical

evidence of bias. Dimensions of methodological quality

associated with estimates of treatment effects in controlled

trials. JAMA 1995;283(5):408–412.

Shikata 2006

Shikata S, Yamagishi H, Taji Y, Shimada T, Noguchi Y.

Single- versus two- layer intestinal anastomosis: a meta-

analysis of randomized controlled trials. BMC Surg 2006;6

(2):1.

Singh 1989

Singh H, Krishnamurthy D, Tayal R, Singh M, Singh K.

Colonic anastomosis in calves: an experimental study. Acta

Vet Hung 1989;37(1-2):167–177.

Thomas 2008

Thomas O, Thabane L, Douketis J, Chu R, Westfall AO,

Allison DB. Industry funding and the reporting quality of

large long-term weight loss trials. Int J Obes (Lond) 2008;

32:1531–1536.

Turnbell 1969

Turnbell RB, Kyle K. Operative Surgery. In: Turnbell RB,

Kyle K editor(s). Operative Surgery. 1. Vol. 5, London:

Butterworths, 1969:663.

Vella 2002

Vella M, ODwyer P. Techniques of bowel resection and

anastomosis. CME J Gynecol Oncol 2002;7(1):290–292.

Walker 2004

Walker KG, Bell SW, Rickard MJ, Mehanna D, Dent OF,

Chapuis PH, Bokey EL. Anastomotic leakage is predictive

of diminished survival after potentially curative resection for

colorectal cancer.. Ann Surg 2004;240(2):255–259.

Whooley 2001

Whooley BP, Law S, Alexandrou A, Murthy SC, Wong

J. Critical appraisal of the significance of intrathoracic

anastomotic leakage after esophagectomy for cancer. Am J

Surg 2001;181(3):198–203.

Zieren 1993

Zieren HU, Muller JM, Pichlmaier H. Prospective

randomized study of one- or two-layer anastomosis following

oesophageal resection and cervical oesophagogastrostomy.

Br J Surg 1993;80(5):608–611.∗ Indicates the major publication for the study



C H A R A C T E R I S T I C S O F S T U D I E S

Characteristics of included studies [ordered by study ID]

Burch 2000

Methods Study design: Prospective randomised controlled trial

Randomization technique: Random permuted blocks of size 10. Opaque sealed en-

velops indicating technique to be used were drawn sequentially

Allocation concealment: Clearly mentioned

Inclusion criteria: Well explained

Exclusion criteria: Well explained

Lost to follow up: Not mentioned

Baseline variables: Matching between both limbs of the trial

Intention-to-treat analysis: Clearly mentioned

Sample size calculations: (power of the study): Not mentioned

Participants Country: United States of America

Number of participants: SGIA: 65 and DGIA: 67

Mean age: SGIA: 44.3 years and DGIA: 44.7 years

M : F = SGIA 42: 23 and DGIA 40 : 27

Interventions SGIA: Continuous 3/0 polypropylene double needle

DGIA: Interrupted 3/0 silk Lembert sutures for outer layer

Continuous 3/0 polyglycolic acid sutures for transmural inner layer

Types of gastrointestinal anastomosis:

• Enteroenterostomy

• Enterocolostomy

• Colocolostomy

Confounding interventions:

• Mechanical bowel preparation and rectal wash

• Prophylactic antibiotics

• Preoperative dietary restrictions

Outcomes Primary: Anastomotic leak

Secondary:

• Duration of anastomosis

• Intra-abdominal abscess formation

• Hospital stay

• Cost analysis

Notes Diagnosis of anastomotic leak was made by using clinical, radiological, and operative

findings

Risk of bias

Bias Authors’ judgement Support for judgement

Random sequence generation (selection

bias)

Low risk Random permuted blocks of size 10.

Opaque sealed envelops indicating tech-

nique to be used were drawn sequentially



Burch 2000 (Continued)

Allocation concealment (selection bias) Low risk Opaque sealed envelops indicating tech-

nique to be used were drawn sequentially

available in operation theatre at the time of

anastomosis

Incomplete outcome data (attrition bias)

All outcomes

Low risk Data reported comprehensively

Selective reporting (reporting bias) Low risk No selective reporting detected

Other bias Unclear risk Rectal anastomosis was excluded

Blinding of participants and personnel

(performance bias)

All outcomes

High risk Not clearly mentioned

Blinding of outcome assessment (detection

bias)

All outcomes

High risk Not clearly mentioned

Everett 1975


Randomization technique: Random selection during laparotomy

Allocation concealment: Not mentioned





Intention-to-treat analysis: Not mentioned


Participants Country: United Kingdom



M : F = SGIA 1: 1.4 and DGIA 1.1 : 1.2

Interventions SGIA: Interrupted 4/0 Supramid full thickness stitches, Gambee type sutures for frontal

layer

DGIA: Continuous 2/0 chromic catgut for all coat stitch

Interrupted Lembert 4/0 supramid sutures for seromuscular layer


• Upper rectal anastomosis (anastomotic line above the peritoneal reflections)

• Lower rectal anastomosis (anastomotic line below the peritoneal reflections)







Everett 1975 (Continued)


Secondary:

• Mortality

• Perioperative complications

Notes • Defunctioning transverse colostomy was preformed in all cases of lower rectal

anastomosis

• Diagnosis of anastomotic leak was made by using clinical, radiological, operative

and postmortem findings

Risk of bias



bias)

Unclear risk Random selection during laparotomy

Allocation concealment (selection bias) High risk Not reported


All outcomes

Low risk All relevant data was reported

Selective reporting (reporting bias) Low risk No selective reporting noticed in the study

Other bias Unclear risk This study encompasses colorectal anasto-

mosis only


(performance bias)

All outcomes

High risk Not reported


bias)

All outcomes




Goligher 1977


Randomization technique: Sequential selection of patients undergoing anterior resec-

tion over a span of

30 months in a major colorectal unit











M : F = SGIA 1.3: 1 and DGIA 1.7 : 1

Interventions SGIA: Interrupted 4/0 Supramid full thickness stitches, Gambee type sutures for frontal

layer

DGIA: Continuous 3/0 chromic catgut for inner layer

Interrupted Lembert 3/0 silk for outer layer


• Upper rectal anastomosis (anastomotic line above the peritoneal reflections)

• Lower rectal anastomosis (anastomotic line below the peritoneal reflection)






Secondary:

• Perioperative complications

Notes • Defunctioning transverse colostomy was performed in all cases of lower rectal

anastomosis

• Diagnosis of anastomotic leak was made by using clinical, radiological, operative

and postmortem findings

Risk of bias



bias)

High risk Sequential selection of patients undergoing

anterior resection

Allocation concealment (selection bias) High risk Not mentioned


All outcomes




Goligher 1977 (Continued)

Selective reporting (reporting bias) Unclear risk No selective reporting noticed in the study

Other bias Unclear risk This study encompasses colorectal anasto-

mosis only


(performance bias)

All outcomes

High risk Not mentioned


bias)

All outcomes

High risk Not mentioned

Irvin 1973


Randomization technique: Random selection for recruitment of the participants


Inclusion criteria: Not explained

Exclusion criteria: Not explained







Mean age: SGIA: 64 13 years and DGIA: 57 19 years

M : F = SGIA 12: 17 and DGIA 17: 14

Interventions SGIA: Interrupted 3/0 silk sero-submucosal, full thickness and Gambee type sutures

DGIA: Continuous full thickness 3/0 chromic catgut for inner layer

Interrupted Lembert 3/0 silk for outer layer


• End to end anastomosis of small intestine

• End to end anastomosis of large intestine including colorectal anastomosis


• Mechanical bowel preparation




Secondary:

• Mortality


findings



Irvin 1973 (Continued)

Risk of bias



bias)

High risk Random selection for recruitment of the

participants without the use of proper ran-

domizations technique



All outcomes


Selective reporting (reporting bias) Low risk No selective reporting noticed in the study

Other bias Low risk None: this study encompasses all levels of

anastomosis in gastrointestinal tract


(performance bias)

All outcomes



bias)

All outcomes


Maurya 1984




Inclusion criteria: Eexplained

Exclusion criteria: Explained





Participants Country: India



M : F = SGIA 36: 24 and DGIA 70 : 42

Interventions SGIA: Interrupted 3/0 silk through and through, full thickness and Gambee type sutures

DGIA: Continuous through and through 3/0 chromic catgut for inner layer interrupted

Lembert 3/0 silk for outer seromuscular layer


• Ileo-ileal anastomosis



Maurya 1984 (Continued)

• Ileo-colic anastomosis

• Colo-colic anastomosis


• Not given..........all acute surgical patients


Secondary:

• Hospital stay

• Bowel sounds

• Passage of flatus

• Duration of intravenous alimentation


findings

Risk of bias



bias)

High risk Random selection of patients

Allocation concealment (selection bias) High risk Not given


All outcomes

High risk All relevant data given

Selective reporting (reporting bias) Low risk All relevant data given

Other bias High risk None


(performance bias)

All outcomes



bias)

All outcomes




Ordorica-Flores 1998


Randomization technique: Patients were randomly divided into the groups assigned

based on a previous list of randomised numbers collected by tables in closed envelopes


Inclusion criteria: Explained

Exclusion criteria: Explained





Participants Country: Mexico



Interventions SGIA: Interrupted 4/0 or 5/0 polyglactin full thickness stitches

DGIA: First layer 4/0 or 5/0 polyglactin Connel-Mayo running stitches and second layer

with separate Lembert stitches using same polyglactin





• Colostomy closure

• Ileostomy closure


• Mechanical bowel preparation




Secondary:

• Hospital stay

• Duration of anastomosis


findings

Risk of bias



bias)

Low risk Good randomisation technique given



All outcomes




Ordorica-Flores 1998 (Continued)

Selective reporting (reporting bias) Low risk No selective reporting detected

Other bias Low risk None


(performance bias)

All outcomes

Low risk Blinding of patient and assessor


bias)

All outcomes


Wayand 1984










Participants Country: Germany



M : F = SGIA 36: 24 and DGIA 70 : 42

Interventions SGIA: Modified Gambee suture using 3/0 or 4/0 polyglactin

DGIA: Sermuscular layer with 3/0 polyglactin and mucosal layer with 2/0 catgut


• Gastric anastomosis





• Not reported


Secondary:

• Mortality

• Peri-operative complications


findings

Risk of bias



Wayand 1984 (Continued)



bias)

Unclear risk Not adequately reported



All outcomes

Unclear risk Not clearly reported

Selective reporting (reporting bias) Unclear risk Not clearly reported

Other bias Unclear risk None: this study encompasses all levels of

anastomosis in gastrointestinal tract


(performance bias)

All outcomes



bias)

All outcomes


Characteristics of excluded studies [ordered by study ID]

Study Reason for exclusion

Carty 1991 Prospective study but without randomizations and trial protocol: comparative trial

Kingsnorth 1989 Prospective study but without randomizations and trial protocol: comparative trial



D A T A A N D A N A L Y S E S

Comparison 1. Anastomosis

Outcome or subgroup titleNo. of

studies

No. of

participants Statistical method Effect size

1 Anastomosis leak 7 842 Odds Ratio (M-H, Fixed, 95% CI) 0.76 [0.44, 1.32]

2 Anastomosis time 2 218 Mean Difference (IV, Fixed, 95% CI) -11.12 [-16.37, -5.

87]

3 Hospital stay 3 390 Mean Difference (IV, Random, 95% CI) -3.08 [-8.49, 2.34]

4 Mortality 4 403 Odds Ratio (M-H, Fixed, 95% CI) 0.56 [0.19, 1.63]

5 Major complications 7 842 Odds Ratio (M-H, Fixed, 95% CI) 0.71 [0.44, 1.16]

6 Anatomotic leak in high quality

trials

3 353 Odds Ratio (M-H, Fixed, 95% CI) 1.13 [0.42, 3.01]

7 Anastomotic leak in poor quality

trials

4 489 Odds Ratio (M-H, Fixed, 95% CI) 0.63 [0.32, 1.24]

Analysis 1.1. Comparison 1 Anastomosis, Outcome 1 Anastomosis leak.

Review: Single layer versus double layer suture anastomosis of the gastrointestinal tract

Comparison: 1 Anastomosis

Outcome: 1 Anastomosis leak

Study or subgroup SGIA DGIA Odds Ratio Weight Odds Ratio

n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI

Burch 2000 2/65 1/67 3.2 % 2.10 [ 0.19, 23.68 ]

Everett 1975 2/40 2/52 5.6 % 1.32 [ 0.18, 9.77 ]

Goligher 1977 5/69 4/66 12.8 % 1.21 [ 0.31, 4.72 ]

Irvin 1973 2/29 3/31 9.1 % 0.69 [ 0.11, 4.47 ]

Maurya 1984 4/60 20/112 44.1 % 0.33 [ 0.11, 1.01 ]

Ordorica-Flores 1998 2/42 3/44 9.5 % 0.68 [ 0.11, 4.31 ]

Wayand 1984 8/103 4/62 15.6 % 1.22 [ 0.35, 4.24 ]

Total (95% CI) 408 434 100.0 % 0.76 [ 0.44, 1.32 ]

Total events: 25 (SGIA), 37 (DGIA)

Heterogeneity: Chi?? = 4.13, df = 6 (P = 0.66); I?? =0.0%

Test for overall effect: Z = 0.98 (P = 0.33)

Test for subgroup differences: Not applicable

0.05 0.2 1 5 20

Favours SGIA Favours DGIA



Analysis 1.2. Comparison 1 Anastomosis, Outcome 2 Anastomosis time.



Outcome: 2 Anastomosis time

Study or subgroup SGIA DGIAMean

Difference WeightMean

Difference

N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI

Burch 2000 65 20.8 (16.9) 67 30.7 (16.9) 82.8 % -9.90 [ -15.67, -4.13 ]

Ordorica-Flores 1998 42 26.17 (29.9) 44 43.16 (29.9) 17.2 % -16.99 [ -29.63, -4.35 ]

Total (95% CI) 107 111 100.0 % -11.12 [ -16.37, -5.87 ]

Heterogeneity: Chi?? = 1.00, df = 1 (P = 0.32); I?? =0%



-20 -10 0 10 20


Analysis 1.3. Comparison 1 Anastomosis, Outcome 3 Hospital stay.



Outcome: 3 Hospital stay

Study or subgroup SGIA DGIAMean

Difference WeightMean

Difference

N Mean(SD) N Mean(SD) IV,Random,95% CI IV,Random,95% CI

Burch 2000 65 7.9 (9.7) 67 9.9 (9.7) 30.9 % -2.00 [ -5.31, 1.31 ]

Maurya 1984 60 11.4 (4.23) 112 18.6 (4.87) 34.2 % -7.20 [ -8.60, -5.80 ]

Ordorica-Flores 1998 42 10.43 (0.5) 44 10.43 (0.5) 34.9 % 0.0 [ -0.21, 0.21 ]

Total (95% CI) 167 223 100.0 % -3.08 [ -8.49, 2.34 ]

Heterogeneity: Tau?? = 21.82; Chi?? = 100.57, df = 2 (P<0.00001); I?? =98%



-10 -5 0 5 10




Analysis 1.4. Comparison 1 Anastomosis, Outcome 4 Mortality.



Outcome: 4 Mortality



Everett 1975 0/40 1/52 14.2 % 0.42 [ 0.02, 10.68 ]

Irvin 1973 2/29 3/31 29.7 % 0.69 [ 0.11, 4.47 ]

Ordorica-Flores 1998 0/42 1/44 16.0 % 0.34 [ 0.01, 8.61 ]

Wayand 1984 3/103 3/62 40.1 % 0.59 [ 0.12, 3.02 ]

Total (95% CI) 214 189 100.0 % 0.56 [ 0.19, 1.63 ]





0.01 0.1 1 10 100




Analysis 1.5. Comparison 1 Anastomosis, Outcome 5 Major complications.



Outcome: 5 Major complications



Burch 2000 4/65 3/67 7.2 % 1.40 [ 0.30, 6.51 ]

Everett 1975 2/40 3/52 6.4 % 0.86 [ 0.14, 5.41 ]

Goligher 1977 5/69 4/66 9.8 % 1.21 [ 0.31, 4.72 ]

Irvin 1973 4/29 6/31 13.0 % 0.67 [ 0.17, 2.65 ]

Maurya 1984 4/60 20/112 33.7 % 0.33 [ 0.11, 1.01 ]

Ordorica-Flores 1998 2/42 4/44 9.6 % 0.50 [ 0.09, 2.89 ]

Wayand 1984 11/103 7/62 20.2 % 0.94 [ 0.34, 2.57 ]

Total (95% CI) 408 434 100.0 % 0.71 [ 0.44, 1.16 ]





0.1 0.2 0.5 1 2 5 10




Analysis 1.6. Comparison 1 Anastomosis, Outcome 6 Anatomotic leak in high quality trials.



Outcome: 6 Anatomotic leak in high quality trials



Burch 2000 2/65 1/67 12.7 % 2.10 [ 0.19, 23.68 ]

Goligher 1977 5/69 4/66 50.3 % 1.21 [ 0.31, 4.72 ]

Ordorica-Flores 1998 2/42 3/44 37.0 % 0.68 [ 0.11, 4.31 ]

Total (95% CI) 176 177 100.0 % 1.13 [ 0.42, 3.01 ]





0.05 0.2 1 5 20


Analysis 1.7. Comparison 1 Anastomosis, Outcome 7 Anastomotic leak in poor quality trials.



Outcome: 7 Anastomotic leak in poor quality trials



Everett 1975 2/40 2/52 7.5 % 1.32 [ 0.18, 9.77 ]

Irvin 1973 2/29 3/31 12.3 % 0.69 [ 0.11, 4.47 ]

Maurya 1984 4/60 20/112 59.2 % 0.33 [ 0.11, 1.01 ]

Wayand 1984 8/103 4/62 21.0 % 1.22 [ 0.35, 4.24 ]

Total (95% CI) 232 257 100.0 % 0.63 [ 0.32, 1.24 ]





0.1 0.2 0.5 1 2 5 10




A P P E N D I C E S

Appendix 1. Search strategy on CENTRAL

We initially searched the The CCCG (Colorectal Cancer Cochrane Group) Controlled Trials Register and the Cochrane Central Register

of Controlled Trials (CENTRAL) on The Cochrane Library (Issue 1, 2011). The following search strategy was used in CENTRAL

and CCCG CTR

1-gastrointestinal anastomosis explode tree 1 (MeSH)

2-oesophageal anastomosis explode all trees (MeSH)

3-gastric anastomosis explode tree 1 (MeSH)

4-small bowel anastomosis explode all trees (MeSH)

5-colonic anastomosis explode tree 1 (MeSH)

6-rectal anastomosis (MeSH)

7-colorectal anastomosis

8-anastomosis*

9-(gastrointestinal continuity*)

10-*bowel resection anastomosis

11-entero=enteric anastomosis

12-(enterocolic anastomosis)

13-colocolic anastomosis

14-(#1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or # 12 or #13)

15- single layer anastomosis(MeSH)

16-double layer anastomosis (MeSH)

17-multiple layer anastomosis (MeSH)

18-(#15 or #16 or #17)

19-anastomotic dehiscence (MeSH)

20-anastomotic leak (MeSH)

21-morbidity of gastrointestinal anastomosis (MeSH)

22-complications of gastroitestinal anastomosis (MeSH)

23-(#19 or #20 or #21 or #22)

24- (#14 or #18 or #23)

We searched the bibliographies of all retrieved and relevant publications identified by these strategies for further studies. We contacted

manufacturers and distributors of suture products as well as relevant government bodies and professional organisations, such as the

Association of Surgeons of Great Britain, Ireland and Associatioon of Coloproctology of Great Britain & Ireland and Euopean Society

of Coloproctology, for details of unpublished and ongoing studies. We did not restrict the inclusion of reports on the basis of language

of publication, date or publication status.

Appendix 2. MEDLINE search strategy

1-exp Gastrointestinal anastomosis/

2-restoration of continuity of gastrointestinal tract .tw.

3-and/1-2

4-(esophageal anastomosis)$.tw.

5-(gastric anastomosis).tw.

6-(small bowel and large bowel anastomosis, colorectal anastomosis*).tw.

7-or/3-6

8-exp single layer anastomosis/

9-exp double layer anastomosis.tw.

10-(multiple layer* adj5 (layered anastomosis* or joint formation*)).tw.

11-exp anastomotic dehiscence/

12-exp anastomotic leak/

13-(complications of anastomosis).tw.

14-or/8-13

15-7 and 14



Appendix 3. EMBASE search strategy

1-exp gastrointestinal anastomosis/

2-(anastomosis adj (gastrointestinal $ or continuation$)).ti,ab.

3-(joint adj5 restoration of gastrointestinal$).ti,ab.

4-or/1-3

5-exp entero-enteric or entero-colic or colo-colic or colorectal anastomosis/

6-single layer anastomosis.ti,ab.

7-(multiple layer $ anastomosis).ti,ab.

8-double layer anastomosis/ti,ab.

9-or 5-8

10- 4 and 11

Appendix 4. Search strategy used in other electronic databases

We used same search strategy as well as MeSH terms to identify target trials in the database of LILACS (The Latin American and

Caribbean Health Sciences Library until April 2011 ) and Science Citation Index Expanded (SCI-E until April 2011).

H I S T O R Y

Protocol first published: Issue 4, 2005

Review first published: Issue 1, 2012

Date Event Description

7 September 2010 New search has been performed This is an updated version of the original protocol published in 2005, with

a new author team

C O N T R I B U T I O N S O F A U T H O R S

MSS: First author of the protocol and review, electronic database search, data extraction, trial scoring, designing of inclusion and

exclusion criteria, data feeding on Excel sheet, data analysis and data interpretation.

MRSS: First co-author of the protocol and review, electronic database search, data extraction, trial scoring, data analysis and data

interpretation

MKB: Second co-author and supervising Consultant Surgeon of the protocol and review, re-checking and confirmation of the extracted

data, conflict resolution if discrepancies in data between authors arise, data analysis, data interpretation and help to write don the

manuscript.



D E C L A R A T I O N S O F I N T E R E S T

All authors confirm that we do not have any potential or actual personal, financial or political interest in this study. We declare that

there was no financial support of any person/company in preparation of this study directly or indirectly.

S O U R C E S O F S U P P O R T

Internal sources

• Authors of this review did not get any internal support in the preparation of this manuscript:, Not specified.

External sources

• Authors of this review did not get any external support in the preparation of this manuscript:, Not specified.

D I F F E R E N C E S B E T W E E N P R O T O C O L A N D R E V I E W

Initially we wanted to include only high quality randomised controlled trials in our review but due to the lack of number of published

trials, we expanded our inclusion criteria and we analysed relatively good and poor quality trials together. We attempted to study if

some difference exists between absorbable or non-absorbable sutures and continuous versus interrupted stitches in making the GIA

measuring their effectiveness in terms of primary and secondary outcomes.There was insufficient data for analysis to reach any reliable

conclusion. In addition we attempted to study different surgical specialties separately involving gastrointestinal anastomosis but it

could not be performed due to lack of data. We also attempted to perform subgroup analyses according to the gastrointestinal segment

included in the trials, type of suture used, as well as type of procedure (elective versus acute) but due to lack of data, it could not be

performed. Re-intervention rate could also not be analysed due to lack of either reporting or investigation.

N O T E S

This is an substantially updated version of a protocol published in 2005, by Alvaro Sanabria et al.

I N D E X T E R M S

Medical Subject Headings (MeSH)

∗Suture Techniques; Anastomosis, Surgical [methods]; Colon [surgery]; Gastrointestinal Tract [∗surgery]; Randomized Controlled

Trials as Topic; Rectum [surgery]; Stomach [surgery]

MeSH check words

Humans



Documents

Cochrane Database of Systematic Reviews (Reviews) || Single layer versus double layer suture anastomosis of the gastrointestinal tract