Intra-peritoneal prophylactic agents for preventing adhesions and adhesive intestinal obstruction after non-gynaecological abdominal surgery

Intra-peritoneal prophylactic agents for preventing adhesions

and adhesive intestinal obstruction after non-gynaecological

abdominal surgery (Review)

Kumar S, Wong PF, Leaper DJ

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

2009, Issue 4

http://www.thecochranelibrary.com

Intra-peritoneal prophylactic agents for preventing adhesions and adhesive intestinal obstruction after non-gynaecological abdominal

surgery (Review)

Copyright © 2009 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 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

12AUTHORS’ CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

13ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

13REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

15CHARACTERISTICS OF STUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22DATA AND ANALYSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Analysis 1.1. Comparison 1 HA/CMC membrane Versus Control, Outcome 1 Incidence of adhesions. . . . . . 23

Analysis 1.2. Comparison 1 HA/CMC membrane Versus Control, Outcome 2 Extent of adhesions. . . . . . . 23

Analysis 1.3. Comparison 1 HA/CMC membrane Versus Control, Outcome 3 Adverse events (Overall). . . . . . 24

Analysis 1.4. Comparison 1 HA/CMC membrane Versus Control, Outcome 4 Wound infection. . . . . . . . 24

Analysis 1.5. Comparison 1 HA/CMC membrane Versus Control, Outcome 5 Intra-abdominal or pelvic infection. . 25

Analysis 1.6. Comparison 1 HA/CMC membrane Versus Control, Outcome 6 Anastomotic leak. . . . . . . . 25

Analysis 1.7. Comparison 1 HA/CMC membrane Versus Control, Outcome 7 Fistula. . . . . . . . . . . . 26

Analysis 1.8. Comparison 1 HA/CMC membrane Versus Control, Outcome 8 Ileus. . . . . . . . . . . . 26

Analysis 1.9. Comparison 1 HA/CMC membrane Versus Control, Outcome 9 First episode of Intestinal obstruction (All

causes). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Analysis 1.10. Comparison 1 HA/CMC membrane Versus Control, Outcome 10 First episode of Intestinal obstruction

(Adhesive) needing surgery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Analysis 2.1. Comparison 2 Ferric Hyaluranate gel Versus Control, Outcome 1 Adverse events (Overall). . . . . 28

Analysis 2.2. Comparison 2 Ferric Hyaluranate gel Versus Control, Outcome 2 Protracted Ileus. . . . . . . . 29

Analysis 2.3. Comparison 2 Ferric Hyaluranate gel Versus Control, Outcome 3 Anastomotic leak. . . . . . . . 29

Analysis 2.4. Comparison 2 Ferric Hyaluranate gel Versus Control, Outcome 4 Wound infection. . . . . . . . 30

Analysis 2.5. Comparison 2 Ferric Hyaluranate gel Versus Control, Outcome 5 Mortality. . . . . . . . . . . 30

30WHAT’S NEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31CONTRIBUTIONS OF AUTHORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31DECLARATIONS OF INTEREST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31DIFFERENCES BETWEEN PROTOCOL AND REVIEW . . . . . . . . . . . . . . . . . . . . .

31INDEX TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

iIntra-peritoneal prophylactic agents for preventing adhesions and adhesive intestinal obstruction after non-gynaecological abdominal

surgery (Review)


[Intervention Review]

Intra-peritoneal prophylactic agents for preventing adhesionsand adhesive intestinal obstruction after non-gynaecologicalabdominal surgery

Senthil Kumar1, Peng F Wong2, David J Leaper3

1Directorate of Surgery, Queens Hospital, Romford, UK. 2Department of Surgery, University Hospital of North Tees, Hardwick, UK.3Department of Wound Healing, Cardiff University, Cardiff, UK

Contact address: Senthil Kumar, Directorate of Surgery, Queens Hospital, Rom Valley way, Romford, Essex, RM7 0AG, UK.

[email protected]. (Editorial group: Cochrane Colorectal Cancer Group.)

Cochrane Database of Systematic Reviews, Issue 4, 2009 (Status in this issue: Unchanged)


DOI: 10.1002/14651858.CD005080.pub2

This version first published online: 21 January 2009 in Issue 1, 2009.

Last assessed as up-to-date: 2 August 2008. (Help document - Dates and Statuses explained)

This record should be cited as: Kumar S, Wong PF, Leaper DJ. Intra-peritoneal prophylactic agents for preventing adhesions and

adhesive intestinal obstruction after non-gynaecological abdominal surgery. Cochrane Database of Systematic Reviews 2009, Issue 1. Art.

No.: CD005080. DOI: 10.1002/14651858.CD005080.pub2.

A B S T R A C T

Background

Intra-abdominal adhesions are common and challenge patients, surgeons and other healthcare providers. They are potentially preventable

and several agents that act as barriers between adjacent peritoneal surfaces have been evaluated for prophylaxis. Efficacy, judged by

systematic reviews, has only been undertaken in gynaecological surgery.

Objectives

To determine efficacy and safety of peritoneal adhesion prophylaxis on incidence, distribution and adhesion-related intestinal obstruction

after non-gynaecological surgery.

Search strategy

The Cochrane Central Register of Controlled Trials, the Cochrane Colorectal Cancer Group specialised register, MEDLINE (1966-

2008), and EMBASE (1971-2008) were searched.

Selection criteria

Blinded and non-blinded, randomised and quasi-randomised clinical trials were considered.

Data collection and analysis

Two authors individually conducted the searches and assessed the quality of studies for inclusion which were analysed using the Revman

Analyses software 5.0.0 provided by the Cochrane collaboration. Meta-analysis used a random effects model.

Main results

Seven randomised trials were eligible; six compared hyaluronic acid/carboxymethyl membrane (HA/CMC) and one 0.5% ferric

hyaluronate gel against controls.

1Intra-peritoneal prophylactic agents for preventing adhesions and adhesive intestinal obstruction after non-gynaecological abdominal

surgery (Review)


mailto:[email protected]

http://www3.interscience.wiley.com/cgi-bin/mrwhome/106568753/DatesStatuses.pdf

juliob

Resaltado

HA/CMC reduced the incidence of adhesions (OR 0.15 (95% CI: 0.05, 0.43); p=0.0005) with reduced extent (WMD -25.9% (95%

CI: -40.56, -11.26); p=0.0005) and severity. There was no reduction of intestinal obstruction needing surgical intervention (odds ratio:

0.84 (95% CI: 0.24, 2.7) with comparable overall morbidity and mortality.

The study of 0.5% ferric hyaluronate gel was prematurely terminated and no valid conclusions could be made but there was a higher

incidence of overall morbidity (OR 5.04; 95% CI: 1.1, 22.9) and ileus (OR: 9.29; 95% CI: 1.57, 54.77; p=0.01).

Authors’ conclusions

Implications for practice

There is evidence that the use of HA/CMC membrane reduces incidence, extent and severity of adhesions which may, theoretically, have

implications in re-operative abdominal surgery. There is no evidence that the incidence of intestinal obstruction or need for operative

intervention is reduced. HA/CMC appears to be safe but there may be a risk of leak when wrapped around an anastomoses. HA/CMC

may be considered for intra-abdominal, adhesion prophylaxis at a surgeon’s discretion and clinical context.

Implications for research

Further research is needed to explore the effectiveness of other agents in abdominal surgery in general. Synergism, using agents which

target different aspects of adhesiogenesis, with exploration effectiveness in a wide range of emergency and elective surgery should be

considered. Longer term outcomes of recurrent intestinal obstruction and chronic pain, identification of high risk groups of patients

with evaluation of cost-effectiveness are required.

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

The use of hyaluronic acid/carboxymethyl cellulose (HA/CMC) membrane, reduces the incidence, extent and severity of

adhesions in the abdomen.

Adhesions in the abdomen cause abnormal bonding between adjacent peritoneal surfaces and are common after operations in the

abdomen. They are composed of fibrous tissue but also contain blood vessels, fat and nerves. They result in a spectrum of problems

that affect the patient (intestinal blockage, infertility and possibly pain); the surgeon (difficulties in access and dissection, prolongation

of operative time, increase in blood loss, predisposition to bowel injury); the health care provider (increased cost due to readmissions

and litigation). Prevention is the key. This review focus on the evaluation of the safety and efficacy of two preventive agents applied

in the abdomen during general surgical operations, Hyaluronic acid /carboxymethyl cellulose membrane and 0.5% ferric hyaluronate

gel.

There is evidence to suggest that use of Hyaluronic acid/carboxymethyl cellulose membrane reduces the incidence, severity and extent

of adhesions.However, it does not reduce the incidence of subsequent intestinal obstruction or need for surgery to treat the obstruction,

when it occurs. It appears to be safe with no significant increase in adverse effects or deaths when compared to control. There is

limited data on 0.5% ferric hyaluronate gel with only one study available. This study did not report on the efficacy of the gel as it was

prematurely terminated because of a significantly higher rate of adverse effects in the patients who were treated with this gel.

B A C K G R O U N D

Intra-abdominal adhesions are usually iatrogenic, affect both gen-

ders and all ages. They may occur in up to 95% of patients who

had undergone a previous laparotomy (Ellis 1997). This contrasts

with a 10% incidence in patients without a prior laparotomy (

Menzies 1990). Pelvic adhesions that follow trans-abdominal gy-

naecological surgery have long been implicated in causing pain

and infertility. Similarly, patients presenting to the general sur-

geon display a spectrum of adhesion related problems or morbid-

ity. The treating surgeon often faces difficulties with access and

obscured normal anatomy which may lead either to an inability

to apply minimal access surgery, prolongation of operative time (

Coleman 2000) or result in potentially serious organ injury such

as inadvertent enterotomy (Van der Krabben 2000). Adhesions are

estimated to account for about a third of all bowel obstructions

and two-thirds of small bowel obstructions in the western world (

Menzies 1992; Menzies 1993).


surgery (Review)


juliob

Resaltado

juliob

Resaltado

Certain surgical operations have been reported to carry a higher

than usual risk of adhesion-related complications. A 25 % inci-

dence of adhesive obstruction has been reported in a series of 1005

ileo-anal pouch procedures (Fazio 1995). Total or subtotal colec-

tomy (Lockhart-Mummery1967; Nieuwenhuijzen 1998) and ap-

pendicectomy (Cox 1993) are other high risk procedures in terms

of their propensity to lead to adhesive intestinal obstruction. In

neonates, surgery for gastroschisis and malrotation are reported to

carry a 15% risk of post-operative adhesive intestinal obstruction (

Wilkins 1986). Managing adhesions and adhesion related compli-

cations is expensive in terms of time (Menzies 2001; Parker 2001),

effort and finances (Ray 1998; Jeekel 1997; Menzies 2001; Wilson

1998) and impacts not only on the patient and the treating physi-

cian but also the health care system at large. The prolongation in

hospital stay, re-admissions, repeated surgical interventions and

litigation impose an enormous burden on the health care systems

globally, making adhesion prevention an important area of health

care intervention research.

As once established, adhesions follow an unpredictable temporal

course, which is attended by a lifetime risk of recurrent symp-

toms, complications, re-admissions and repeat surgery, prevention

seems to be the most viable strategy yet. Peritoneal injury, the at-

tendant inflammation, imbalance in plasmin system and proxim-

ity of injured surfaces are the crucial elements in adhesion for-

mation. Hence, preventive strategies have been designed targeting

these steps individually or in combination. Injury limitation, a vi-

tal aspect of prophylaxis is largely a surgical endeavour achieved by

meticulous attention to operative technique and tissue handling.

However, as injury can only be minimised and not completely

abolished, adjuncts to injury limitation techniques are necessary to

reduce the risk of adhesions. Systemic and intraperitoneal pharma-

cological strategies like anti-inflammatories, anticoagulants, agents

targeting the plasmin cascade and hormones have not met with

much success clinically (diZerega 1994; Risberg 1997).

A number of fluid and solid phase (membranes, sheets) agents

that primarily act as barriers to contact between two adjacent peri-

toneal surfaces, have been evaluated clinically for their potential

for adhesion prophylaxis. Examples of fluid agents include 4%

Icodextrin (Adept®, ML laboratories plc.,UK) and Hyaluronic

acid-phosphate buffered saline (Sepracoat®, Genzyme Corp.,

USA). Examples of solid phase agents include, Hyaluronic

acid /Carboxymethyl cellulose membrane (Seprafilm®, Genzyme

Corp., USA), Oxidised regenerated cellulose (Interceed®, John-

son&Johnson medical Inc., USA), and expanded Polytetraflu-

oroethylene (Gore-tex®, WL Gore& Associates, USA ). These

agents differ in their chemical composition, biodegradability, tis-

sue interactions and hence the adverse effect profiles.

An intervention such as adhesion prophylaxis, that is likely to have

a wide impact on day to day surgical practice and on the health

care economics needs to be scrutinised for safety and efficacy with

specific reference to the context in which it is used and in relation

to relevant outcomes.Barrier and non-barrier agents in prophylaxis

of adhesions have been evaluated for efficacy by systematic reviews,

but only in the context of gynaecological surgery (diZerega 2002),

usually with respect to their effect on fertility rates and pain (

Ahmad 2008, Metwally 2006).

The purpose of the current review is to provide an evidence based

appraisal of the safety and efficacy of intra-peritoneal agents in

general surgical operations from randomised controlled trials that

have assessed adhesion reformation (severity, distribution and fre-

quency) and /or adhesion related intestinal obstruction as the main

outcome measure.

O B J E C T I V E S

Primary objective:

To determine the efficacy of intra-peritoneal prophylactic agents

in reducing the incidence, distribution and severity of adhesions

in non-gynaecological abdominal surgery.

Secondary objectives:

To determine the effect of intra-peritoneal prophylactic agents on

the incidence and outcome of adhesion related intestinal obstruc-

tion.

To determine the safety and adverse effect profile of intra-peri-

toneal prophylactic agents in non-gynaecological intra-abdominal

surgery.

To detect any changes in health related quality of life from the use

of intra-peritoneal prophylactic agents in this subset of patients.

M E T H O D S

Criteria for considering studies for this review

Types of studies

Inclusion criteria:

Randomised and Quasi-randomised clinical trials addressing the

safety and/or efficacy of prophylactic intraperitoneal agents in pre-

vention of adhesions and adhesion related complications were in-

cluded.

Both blinded and non-blinded studies were considered.

Only parallel group and factorial designs were considered.

Only studies involving non-gynaecological abdominal surgery in-

volving the peritoneal contents were evaluated for inclusion.

In studies reporting on efficacy, the patients required to have had

the intra-peritoneal adhesions assessed at a second operation -

either open or laparoscopic.


surgery (Review)


Types of participants

Inclusion criteria:

Patients of all ages and both genders were included.

Patients who underwent a non-gynaecological abdominal surgery

either open or laparoscopic, involving the use of a prophylactic

intra-peritoneal agent for adhesion prevention, during the course

of the surgery.

Exclusion criteria:

Patients undergoing primarily a transabdominal gynaecological

surgery .This would be defined for the purpose of this review as

any transabdominal primary operative intervention on the female

reproductive tract. However, patients who had a primary general

surgical operation but also had to undergo a surgical procedure on

the female reproductive tract during the course of the operation

were considered to be included in the review.

Types of interventions

Intra-operative intra-peritoneal incorporation or instillation of a

solid phase agent (membranes or sheets) or a fluid agent respec-

tively, with a view to prevent intra-abdominal adhesions and ad-

hesion related complications.

The prophylactic agent studied should have been compared with

either no intervention or against a placebo or against another pro-

phylactic agent.

Both open and minimal access surgeries involving the intraperi-

toneal contents and employing the interventions were considered.

There was no restriction as to the grade or level of experience of

the operating surgeon performing the interventions.

Types of outcome measures

Inclusion criteria:

Primary outcome measure:

1.The incidence, distribution and severity of adhesions assessed

during the course of a second open surgical or laparoscopic oper-

ation

Secondary outcome measures:

1.The incidence and outcome of adhesion-related intestinal ob-

struction.

2.Incidence and outcome of adhesion related re-admissions

3.Local, regional and systemic complications arising from the use

of fluid and membrane agents

4.Patient reported health related quality of life recorded by a vali-

dated generic or specific quality of life assessment tool.

Exclusion criteria for outcome measures:

1. Where the main outcome study of the study was pain

2. Where the main outcome measure of the study was fertility

Search methods for identification of studies

The following bibliographic electronic medical databases were

searched for publications addressing the above clinical problem

1.The Cochrane Central Register of Controlled Trials (CEN-

TRAL)

2.The Cochrane Colorectal Cancer Group specialised register (SR-

COLOCA)

3.MEDLINE from 1966 to 2008 (April)

4.EMBASE from 1971 to 2008 (April)

Two authors individually conducted the searches.The following

search strings were used where appropriate, without language re-

strictions.

Strategy A:

1.Adhesi$ - LIMIT to human studies

2.Abdo$ OR Intraabdominal

3.Peritoneal OR Intraperitoneal

4.COMBINE 2 OR 3

5.COMBINE 1AND 4

6.Prevent$ OR Prophyla$

7.COMBINE 5 AND 6

The titles and abstracts when available were scrutinised to select

relevant controlled studies addressing the safety and efficacy of the

use of these agents in non-gynaecological abdominal surgery.

Strategy B:

The following text words were used individually to identify studies

reporting results from the specific use of the respective products

1.“SEPRAFILM”

2.“SEPRACOAT”

3.“INTERCEED”

4.“ADEPT”

5.“ICODEXTRIN”

6.“GORETEX”

7. “SURGIWRAP”

8. “DEXTRAN”

9. “HYSCON”

10. “CARBOXYMETHYLCELLULOSE”

11. “ HYALURONAN”

12. “ HYALURONIC ACID”

13. “ POLYTETRAFLUOROETHYLENE”

14. “ OXIDISED REGENERATED CELLULOSE”

15. “ PHOSPHATIDYLCHOLINE”

16.COMBINE 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 OR 14 OR 15.

17.Adhesi$

18.COMBINE 16 AND 17

The search was limited to human studies




Strategy C:

1.Adhes$ LIMIT to human studies

2.Intestinal obstruction OR Bowel obstruction

3.COMBINE 1 and 2

4.Complications

5.COMBINE 1 and 4


surgery (Review)


6 COMBINE 3 OR 5, avoid duplicates

7.Prevent$ OR Prophyl$

8.COMBINE 6 AND 7.




Reference lists of retrieved full text articles were searched for rele-

vant additional studies.

The manufacturers of the prophylactic agents mentioned above

were contacted when necessary to acquire information about any

published trials known to them that could be of relevance.

The health technology assessment (HTA) database was also

searched for relevant publications.

Data collection and analysis

STUDY SELECTION

Two reviewers (SK and PFW) conducted a methodical search of

the electronic bibliographic medical databases and other sources as

outlined in the search strategy. Trials were considered for inclusion

in the review if they fulfilled the following criteria:

1. Prospective randomised or quasi-randomised controlled trials

in humans.

2.The procedure performed was an intra-peritoneal non-gynaeco-

logical surgery, either open or laparoscopic.

3.The trial compared an agent against either no treatment or an

appropriate placebo or against another prophylactic agent.

4.Trial addressed the safety and/or efficacy of the intra-peritoneal

prophylactic agents.

After initial screening of the title and abstract, the full-text article

of the studies were retrieved in the following situations:

1.Study fulfils criteria on initial screening of the abstract and title

2.Study may potentially fulfil criteria but there is not enough in-

formation in the abstract alone to justify inclusion.

3.No abstract is available but the title suggests a potential that the

study may fulfil criteria.

The full text article was then scrutinised independently by two re-

viewers to confirm that the study justified inclusion. A specifically

designed structured flow chart with a set of check lists incorpo-

rated was used in the selection procedure.

QUALITY ASSESSMENT:

Quality assessment was by a component based system incorporat-

ing aspects of internal validity outlined below. Composite scoring

systems for quality assessment were not used.The domains consid-

ered towards quality assessment, were selected to detect any poten-

tial for selection bias, detection bias and attrition bias. The studies

were scrutinised for details regarding the following components

of quality:

1. Randomisation:

Good quality: True randomisation ( Computer generated random

number sequences, sequences from validated random number ta-

bles, coin tossing, casting dice, shuffling cards or similar methods

)

Poor quality: Quasi-randomisation (based on date of birth, hos-

pital number, date of admission, alternating sequence or similar

methods ).

2.Allocation concealment:

Good quality: sequentially labelled sealed, opaque envelopes; Cen-

tralised randomisation revealed by telephoning before assignment

or similar methods.

Poor quality: unsealed or non-opaque envelopes, open allocation

schedule, all procedures based on quasi-randomisation.

3.Blinding:

Good quality: Assessor or observer is blinded to the intervention

that the patient has had.

Poor quality: Assessor not blinded; Only the patient and or the

surgeon applying the intervention is blinded ; No blinding used.

4.Handling of attrition:

Good quality: Analysis on an intention to treat basis; patients lost

to follow-up, drop outs, protocol violations reported and causes

discussed; proportion of patient attrition < 15% (Chalmers 1981).

Poor quality: Participant attrition poorly described or analysed.

Attrition rate > 15% of total recruited.

Studies were assigned an overall grade as follows: A (All criteria for

good quality are met, low risk of bias; would be classified as a good

quality study), B (One or more criteria of good quality only partly

met, moderate risk of bias; would be classified as an intermediate

quality study), C (One or more criteria for good quality not met,

high risk of bias; would be classified as a poor quality study).

The grading however was not used to weight the summary out-

comes in the meta-analysis.

DATA COLLECTION :

Data extraction from the studies was standardised by specially

designed forms and was double checked by a second reviewer

independently. It was not possible to blind the reviewers to the

source of the articles (authors, institution, journal, funding). The

data collected was in relation to the following characteristics:

Methods:

-Setting: Geographical location; single centre or multicenter

-Type of randomisation: True, Quasi-.

-Method of randomisation : Generation of sequence; allocation

concealment.

-Extent of blinding: Patient; surgeon employing the intervention;

observer; non-blinded; unclear; not stated.

-Power calculations: Not applicable; not stated; unclear; provided;

if provided then statement of power.

-Size of the study: Total patients recruited; total included in anal-

ysis; whether size appropriate to a priori statement of study size.

-Methods employed to deal with attrition and missing data

-Type of adhesion assessment: second look laparoscopy; second


surgery (Review)


look laparotomy.

-Time of adhesion assessment in relation to the index surgery.

-Duration of the study.

-Duration of follow-up.

Participants:

-Age and gender distribution.

-Type of underlying abdominal pathology (peritonitis or not; be-

nign or malignant; Inflammatory or non-inflammatory bowel

pathology ; mixed, unclear; not stated).

Intervention:

-Setting: Type of surgery- laparoscopic or open; emergency or

elective; mixed.

-Agent(s) used and comparator: agent Vs none; agent Vs placebo;

agent Vs agent.

-Dose or size of agent used.

-Area of application: General peritoneal cavity; all surgical sites;

identified raw areas or peritoneal defects; around anastomosis; un-

der the laparotomy wound; unclear; not stated.

Outcome measures and results:

1.Incidence of adhesions

2.Distribution/ extent of adhesions.

3.Severity/type of adhesions

4..Incidence of morbidity attributable to adhesions or interven-

tion.

5.Incidence of mortality attributable to adhesion or intervention.

6.Incidence of re-admissions

7.Incidence of clinical bowel obstruction.

8.Incidence of operative intervention attributable to adhesions.

9.Quality of life scores

DATA ANALYSIS AND SYNTHESIS:

The statistical package Metaview of Revman 4.2.6, provided by the

Cochrane collaboration was used to analyse and synthesise data.

For dichotomous data such as presence or absence of adhesions,

incidence of morbidity and mortality attributable to adhesions,

incidence of intestinal obstruction and incidence of re-admissions

, the impact of the intervention is expressed as an odds ratio with

95% confidence intervals. For continuous data such as extent of

adhesions measured as absolute area (square millimetres or square

centimetres) or as a proportion or percentage of wound area, the

effect size was estimated by computing the weighted mean differ-

ence with 95% confidence intervals.

The tables of comparison are for the following outcomes:

1.Incidence of adhesions

2.Extent of adhesions

3.Incidence of morbidity (overall and specific)

4.Incidence of mortality

5.Incidence of intestinal obstruction (overall and those needing

surgical intervention)

A meta-analysis was performed where appropriate. There were

too few studies (7) satisfying the inclusion criteria to explore the

potential for publication bias by a funnel plot of sample size plotted

against the odds ratio of primary outcome. Statistical heterogeneity

was tested by inspection of the forest plot and computation of the

chi square statistic for heterogeneity with the probability of a type

I error set at 10% ( P value <0.1 considered significant). Meta-

analysis when appropriate, was performed assuming a random

effects model (Dersimonian and Laird method).

R E S U L T S

Description of studies

See: Characteristics of included studies; Characteristics of excluded

studies.

NUMBER AND TYPES OF STUDIES

A total of 12 studies reporting on the use of intra-abdominal agents

for adhesion prophylaxis in general surgical operations were iden-

tified, of which 5 were excluded. Four of the five excluded studies

were excluded because they were non-randomised clinical studies

(Kudo 2004; Arnbjornsson 1983; Wang 1994; Salum 2001). One

other study (Beck 1997) was excluded because it was a dual pub-

lication of identical data previously published (Becker 1996). All

7 included studies were prospective randomised controlled trials.

However, there were 2 publications (Beck 2003; Fazio 2006), from

one clinical trial reporting on different outcomes which meant

that there were a total of 8 publications from 7 clinical trials

TYPES OF INTERVENTION

Six of the seven

studies involved Hyaluronic acid/Carboxymethylcellulose mem-

brane (Seprafilm) as the intervention agent against a control group

which did not receive any specific anti-adhesion prophylaxis (Beck

2003; Becker 1996; Cohen 2005; Fazio 2006; Kusunoki 2005;

Tang 2003; Vrijiland 2002). The only other study used 0.5% Fer-

ric hyaluranate gel (Intergel) as adhesion prophylaxis in the inter-

vention group while the control group received distilled water (

Tang 2006).

NUMBER OF PATIENTS

There was a wide variation among the studies in the total num-

bers of patients recruited. This ranged from only 32 patients in

one study (Tang 2006) to 1791 patients in another (Beck 2003).

Only two of the other five studies, recruited more than a hundred

patients (Becker 1996; Cohen 2005).

TYPES OF PATIENTS

All studies involved adult patients of both genders and excluded

pregnant patients. Three studies recruited patients who underwent

a wide range of abdominopelvic resective surgery (involving small

bowel, large bowel and rectum) (Beck 2003; Fazio 2006; Tang

2006) while the others included only a more limited group of pa-

tients defined on the basis of the underlying pathology or type of

surgery. Two studies limited their recruitment to patients undergo-

ing colectomy and ileoanal pouch procedure for ulcerative colitis


surgery (Review)


or familial adenomatous polyposis (Becker 1996; Cohen 2005).

One study included only patients who had rectal cancer who un-

derwent preoperative radiotherapy followed by staged surgery (

Kusunoki 2005). Another study recruited only patients who un-

derwent a rectal resection and needed an ileostomy (Tang 2003)

while one restricted recruitment to patients undergoing Hartmann

resection for sigmoid diverticulitis or obstructing recto-sigmoid

lesions (Vrijiland 2002).

TYPES OF OUTCOMES

Five out of the 8 included publications reported on outcomes

which included aspects of both efficacy and safety (Becker 1996;

Cohen 2005; Tang 2003; Tang 2006; Vrijiland 2002) while two

studies reported on efficacy alone (Fazio 2006; Kusunoki 2005)

and one publication on safety alone (Beck 2003). Efficacy of pro-

phylactic agents was reported in 7 trials. The relevant outcome

measures studied were their effect on the incidence of adhesions

in 3 publications (Becker 1996; Vrijiland 2002; Cohen 2005), ex-

tent of adhesions in 3 publications (Becker 1996; Cohen 2005;

Kusunoki 2005), the severity of adhesions in 5 publications (

Becker 1996; Cohen 2005; Kusunoki 2005; Tang 2003; Vrijiland

2002), incidence of bowel obstruction in 2 publications (Fazio

2006; Tang 2006) and their effect on the quality of life in one

study (Tang 2006).

The safety profile of the agents used were reported in 6 publi-

cations. The relevant outcome measures studied were mortality,

overall morbidity and specific morbidity such as wound infection,

anastomotic leak, fistula and intra-abdominal abscess.

Risk of bias in included studies

Of the 7 prospective randomised controlled trials included, 5 were

multi centre trials involving eight (Vrijiland 2002), eleven (Becker

1996), twelve (Cohen 2005), sixteen (Beck 2003) and twenty two

(Fazio 2006) centres. Three trials were based in single institutions

(Tang 2003; Kusunoki 2005; Tang 2006). Sample size calculation

was described in 4 studies (Beck 2003; Becker 1996; Fazio 2006;

Tang 2006). Generation of the randomisation sequence was by

using standard random number tables in one study (Beck 2003;

Fazio 2006) and by a computer generated sequence in 2 studies

(Vrijiland 2002; Tang 2006). In 4 studies the exact method of

generation of the random sequence was not mentioned (Becker

1996; Cohen 2005; Kusunoki 2005; Tang 2006). Allocation con-

cealment was by the use of sealed envelopes in 3 studies (Vrijiland

2002; Tang 2003; Kusunoki 2005), and by resorting to telephon-

ing a remote point of central randomisation at the time of inter-

vention in 1 study (Tang 2006).

Allocation concealment was not clearly evident in 4 studies (Beck

2003; Fazio 2006; Becker 1996; Cohen 2005). Blinded outcome

evaluation to minimise bias was described in 5 studies (Beck 2003;

Fazio 2006; Becker 1996; Cohen 2005; Tang 2006).

Only five studies prospectively defined at least some of the key

outcomes (Fazio 2006; Kusunoki 2005; Tang 2003; Tang 2006;

Vrijiland 2002). Analysis was on an intention to treat basis in 3

studies (Vrijiland 2002; Cohen 2005; Tang 2006).

The period of follow up varied depending on the outcome re-

ported. For example, studies reporting on safety related outcomes

such as morbidity had a shorter follow up than those reporting

on efficacy related outcomes such as the incidence of adhesions

or bowel obstruction. The period of follow up was clearly stated

in 4 publications - Beck 2003 (1 month follow up for adverse

events); Tang 2003 (median follow up of 22.8 months in phase I

and 9.6 months in phase II for adverse events); Fazio 2006 (Mean

follow up of 3.5 years for intestinal obstruction), and Tang 2006

(1 year follow up for intestinal obstruction). The follow up period

for safety related outcomes was not clearly evident in 3 studies (

Becker 1996; Cohen 2005; Vrijiland 2002). In studies reporting

on the efficacy of agents in preventing the incidence and or severity

of adhesions, there was a wide variation in the duration between

the index surgery and the second surgery used for assessment of

adhesions. In Becker 1996, the duration was from 8-12 weeks;

in Cohen 2005 it was 8-14 weeks; Kusunoki 2005 states that the

second surgery was performed at least 3 months after the index

surgery, but does not provide further details; Vrijiland 2002 re-

ports a median duration of 5 months in the intervention group

and 4 months in the control group between index surgery and

assessment; Tang 2003 report two different phases in the trial.

In phase I, the intervention group had ileostomy closure at three

weeks from index surgery, while the control group had the closure

performed at 6-12 weeks. In phase II both the intervention and

control groups had ileostomy closure at 3 weeks.

Attrition of patients either due to exclusion after recruitment or

loss to follow up had a wide variation depending on the outcome

being studied and the duration of follow up. Three studies had no

attrition and were able to report outcomes on all recruited patients

(Beck 2003; Cohen 2005; Fazio 2006). Two studies had low attri-

tion rates of 4.4% (Becker 1996) and 4.8% (Kusunoki 2005). The

other three studies had a much higher degree of attrition, Tang

2003 (38%), Vrijiland 2002 (41%) and Tang 2006 (12.5% at 3

months and 62.5% at 1 year). In one study (Fazio 2006), 90 pa-

tients were excluded post-randomisation, from the final analysis.

The excluded patients were analysed as a different subgroup and

results reported separately from the main cohort of patients in the

study. The reason stated is that these 90 patients, had adhesive

intestinal obstruction, at recruitment while the others did not.

However exclusion after randomisation is likely to introduce bias

and as shown in the analysis, including this subgroup of patients

in the meta-analysis significantly affects the point estimate of the

effect (Analysis 1.10).

Overall 5 studies could be described as grade B studies (interme-

diate quality) (Beck 2003; Fazio 2006; Becker 1996; Cohen 2005;

Kusunoki 2005). Three other studies were assigned grade C (poor

quality) mainly because of a high attrition rate (Vrijiland 2002;

Tang 2003; Tang 2006).


surgery (Review)


Effects of interventions

The results may be summarised in two broad categories: those

relating to the efficacy of the agents and those pertaining to their

safety.

A. EFFICACY:

[1] INCIDENCE OF ADHESIONS

Data was available for the use of only one agent- Hyaluronic

acid/carboxymethyl cellulose membrane. The only study which

used 0.5% Ferric hyaluronate gel did not report on its effect on

the incidence of adhesions (Tang 2006).

Hyaluronic acid/carboxymethylcellulose membrane

(seprafilm) versus control:

Three studies reported on the effect of use of Hyaluronic

acid/carboxymethyl cellulose membrane on the incidence of ad-

hesions (Becker 1996; Vrijiland 2002; Cohen 2005). There were

a total of 97 events amongst 164 patients in the treatment group

compared to 158 events amongst 171 patients in the control arm.

The use of hyaluronic acid/carboxymethyl cellulose membrane re-

duced the incidence of adhesions significantly [OR 0.15 (95% CI:

0.05, 0.43); p=0.0005].

[2] EXTENT OF ADHESIONS

Three studies using Hyaluronic acid/carboxymethylcellulose

membrane as the interventional agent, reported on the extent of

adhesions as a relative measure such as the proportion of either

the main wound length or peristomal area involved by adhesions (

Becker 1996; Fazio 2006; Kusunoki 2005). In two studies the ex-

tent of adhesions was measured on an interval scale as a percentage

of the main laparotomy wound and data analysed comparing dif-

ferences between means or median among the two groups (Becker

1996; Cohen 2005). Becker and colleagues found the mean extent

of adhesions in the intervention group (n=85) to be 23% (s.d. 34)

compared to 63% (s.d. 34) in the control group (n=90) [ p<0.01;

student’s t test]. Cohen and colleagues, reported a mean of 31%

involvement of the main laparotomy wound with adhesions in

the intervention group compared to 65% in the control group.

The mean distribution and error bars representing standard de-

viation were provided in this publication. The measured differ-

ences between the two groups however, were analysed using a non-

parametric method and are reported to be statistically significant

[p<0.001; Mann-Whitney test].

Data from these two studies were used in the meta-analysis. The

use of hyaluronic acid/carboxymethyl cellulose membrane signif-

icantly reduced the extent of adhesions to the midline wound

[WMD -25.9%(95%CI: -40.56, -11.26); p=0.0005]

In one other study (Kusunoki 2005), the adhesions were measured

on an ordinal scale with the extent of adhesions in individual

patients given a grading of 0 (no adhesions), 1 (covering up to 25%

of the length of the wound or peristomal area), 2 (covering 25-

50% of the length of the wound or peristomal area), 3 (covering

more than 50% of the length of the wound or peristomal area).

The extent of midline and peristomal adhesions were measured

in 30 patients in the intervention group and 29 patients in the

control group. When midline adhesions were compared, 26 out

of 30 patients in the intervention group had a grade of 0, while

15 out of 29 patients in control group had grade 2 adhesions.

The differences between the groups were statistically significant

(p<0.01; Mann-Whitney test). Comparing peristomal adhesions,

18 of 30 patients in the interventional group had grade 1 adhesions

while 23 out of 29 patients in the control group had grade 2 or

3 adhesions. The differences between the groups was statistically

significant (p ,0.01; Mann-Whitney test).

[3] SEVERITY OF ADHESIONS

Five studies using Hyaluronic acid/carboxymethylcellulose mem-

brane as the interventional agent reported on the severity of adhe-

sions (Becker 1996; Vrijiland 2002; Tang 2003; Kusunoki 2005;

Cohen 2005). There was a wide variation in the methods and def-

initions used to assess severity which precluded meta-analysis.

Becker 1996 used a three point grading system (grades 1-3) which

assessed thickness and vascularity . There were 90 patients in

the control group and 85 in the intervention group. The use

of Hyaluronic acid/carboxymethylcellulose membrane was asso-

ciated with significantly less severe adhesions as assessed by their

grading system (p<0.01; Wilcoxon rank sum test).

Vrijiland 2002 used the product of an estimate of severity (which

was graded from 1-4) and the distribution of adhesions to give a

composite score which was used as the comparative measure. The

median composite severity score for the whole of the midline in-

cision was significantly lower in the intervention group (18) when

compared to the control group (50) [p=0.002 ; Mann-Whitney

test].

Tang 2003 classified peristomal adhesions by severity into 3 groups

(grades 1-3). Each of the 4 quadrants around the stoma was scored

separately and a total score was calculated by summation. In Phase

one of the trial, assessment was performed 3 weeks after index

surgery in the intervention group (n=26) while the assessment

in the control group was at 6 weeks or after (n=39). There were

no differences in the mean adhesion severity scores between the

two groups in this phase (p=0.84; student’s t test). In the second

phase of the study, a subsequently recruited cohort of patients were

assessed at 3 weeks after index surgery in both groups (n= 21 in the

intervention group and n= 22 in the control group). The mean

adhesion severity score was significantly less in the intervention

group (p=0.02; student’s t test).

Kusunoki 2005 graded the severity of adhesions based on thick-

ness an vascularity into four groups (grades 0-3). There were 30

patients in the intervention group and 29 patients in the control

group. Both midline adhesions and peristomal adhesions were sig-

nificantly less severe in the intervention group (p<0.01; Mann-

Whitney test).

Cohen 2005 used a four point grading system (grades 0-3). There

were 58 patients in the intervention group and 60 patients in the

control group. In the intervention group, 19 patients had grade 2


surgery (Review)


adhesions and 6 patients had grade 3 adhesions compared to 25

and 21 patients respectively in the control group. These differences

in severity were reported to be statistically significant ( p<0.001;

Wilcoxon rank sum test).

[4] INCIDENCE OF RE-ADMISSIONS

None of the included studies had provided specific data on the

incidence of re-admissions in the patients studied.

[5] INCIDENCE OF BOWEL OBSTRUCTION AND OPER-

ATIVE INTERVENTION FOR BOWEL OBSTRUCTION

Overall three out of the eight publications reported on the inci-

dence of bowel obstruction as an outcome measure.

Hyaluronic acid/carboxymethylcellulose mem-

brane (seprafilm) versus control: Two studies reported on the

incidence of intestinal obstruction (Kusunoki 2005; Fazio 2006).

Kusunoki 2005 followed up their patients for 24-66 months (me-

dian 43.6 months) while Fazio 2006 reported a follow up of be-

tween 24-60 months (mean 42 months). Two types of incidences

were reported. All causes bowel obstruction includes a diagnosis of

bowel obstruction in patients in the follow up period, arrived on

the basis of clinical, radiological and/ or operative criteria. Bowel

obstruction attributable to adhesions includes only the subset of

patients who either required a laparotomy or laparoscopy to diag-

nose or treat bowel obstruction and hence had objective operative

evidence to support the adhesive aetiology of obstruction. Fazio

2006 presented their results in two subgroups. Ninety patients

who had adhesive intestinal obstruction at recruitment have been

analysed as a separate subgroup - with a higher incidence of ad-

hesive obstruction in the treatment group (8/42) compared to the

control group (4/48). This essentially means a post-randomisation

exclusion of 90 patients from the initial recruitment cohort of pa-

tients. The authors of this review, feel that such post randomisa-

tion exclusion is inappropriate and introduces a significant bias to

the results (as borne out by the forest plot 1.10.1, which clearly

illustrates the effect of excluding and including this subgroup of

90 patients in the meta-analysis). For the purposes of this review,

the authors have included the data on the subgroup of 90 patients

in the meta-analysis.

The total incidence of intestinal obstruction of all causes was 111

out of 912 patients in the intervention group compared to 111 out

of 938 patients in the control group. The differences between the

two groups was not statistically significant [OR: 0.84 (95% CI:

0.34, 2.08)]. Similarly, in the subgroup of adhesive intestinal ob-

struction needing surgical intervention, the incidence of intestinal

obstruction in the treatment group (24 out of 914) and the control

group (36 out of 939) were not statistically significant [OR: 0.84

(95% CI: 0.24, 2.7)].

0.5% Ferric Hyaluronate gel: One study reported the incidence

of intestinal obstruction during follow up (Tang 2006). The study

was underpowered due to premature termination. Even amongst

the recruited patients, the attrition rate in the study for evaluation

of this particular outcome measure was unacceptably high. Of the

total of 32 patients recruited in the study only 20 patients (10

patients in each arm of the study) completed 1 year follow up for

evaluation of the incidence of intestinal obstruction. One patient

in the treatment group was found to satisfy the criteria set out in

the study, for intestinal obstruction during the one year of follow

up reported. So no valid conclusions can be drawn from this study

on the effect of adhesion prophylaxis on the incidence of adhesive

obstruction.

[6] INCIDENCE OF OPERATIVE INTERVENTION OF

OTHER CAUSES, ATTRIBUTABLE TO ADHESIONS

None of the included studies had provided data specifically relat-

ing to adhesion related operative intervention apart from those

presenting with intestinal obstruction discussed above. Though

one study set out to evaluate the need for adhesiolysis during fol-

low up, as an outcome measure (Tang 2006), this did not mate-

rialise as the study was terminated prematurely. None of the 20

patients followed up for 1 year needed adhesiolysis. In summary,

the recruitment was small (total of 32 patients), follow up short

(1 year), attrition rate for follow up high (37.5%) and the event

rate too low (zero) to draw any valid conclusions.

[7] QUALITY OF LIFE MEASURES

Quality of life is an important patient reported outcome increas-

ingly used to assess effectivenes of interventions. However, it would

be a difficult one to measure in the context of adhesions due to a

number of confounding factors. Only one trial attempted to study

the effect of adhesion prophylaxis on the quality of life (Tang 2006)

using the Gastrointestinal Quality Of Life Index tool. Though

baseline measurements were available for 29 patients (n=15 for

treatment group and n=14 for control group), the attrition rate

was high with only a total of 11 patients ( n=5 for treatment group

and n=6 for control group) completing the quality of life ques-

tionnaire at the end of 1 year. Moreover, only 3 patients completed

all three sets of questionnaires (i.e. baseline, 6 months and 1 year).

Because of the high attrition rate and the non-availability of indi-

vidual patient data, no definite conclusions can be drawn on the

effect of the intervention on the quality of life.

B. SAFETY

Six studies reported on at least one aspect of safety. Five of these

studies used Hyaluronic acid / carboxymethyl cellulose and the

results were reported at the end of the study period as planned. One

study evaluated the safety of 0.5% Hyaluranate gel against placebo

(Tang 2006). Though the study was intended to recruit about

700 patients, the high incidence of morbidity in the intervention

group resulted in a premature termination of the trial within 6

weeks of commencement after recruitment of only 32 patients.

[1] INCIDENCE OF OVERALL MORBIDITY

A total of 4 studies reported on overall adverse events seen in the

intervention and the control groups. The specific list of compli-

cations for which the patients were evaluated for, varied between

different studies, but was applied equally between the two arms


surgery (Review)


within a study. The overall morbidity included both abdominal

(such as abdominal abscess, anastomotic leak) as well as extra-ab-

dominal (such as fever, pulmonary embolism) adverse events.

Hyaluranate/CMC membrane: Three studies recruiting a total of

2094 patients, were available for evaluation of safety of HA/CMC

membrane. The total event rate in all three studies together was

relatively high at 759 (36.2%). There was no significant clinical

or statistical heterogeneity amongst the studies. There was a total

of 390 events amongst 1032 patients in the intervention group

and a total of 369 events in 1062 patients in the control group.

The incidence of overall morbidity was comparable between the

intervention and control groups [OR: 1.19; 95% CI: 0.97, 1.46;

p=0.10].

0.5% Ferric Hyaluranate gel: one study evaluated the safety of

0.5% hyaluranate gel against placebo (Tang 2006). There were a

total of 32 patients of which 17 were in the intervention group. At

30 day follow up, complications were 5 times more common in

the intervention group when compared to the control group [OR

5.04; 95%CI:1.1; 22.9]. This resulted in the premature termina-

tion of the study, within 6 weeks of initiation.

[2] INCIDENCE OF ABDOMINAL OR PELVIC INFEC-

TION

Hyaluranate/CMC membrane: Three studies reported on the

incidence of post-operative intra-abdominal or pelvic infection

after the use of Hyaluranate/CMC membrane (Becker 1996; Beck

2003; Cohen 2005). A total of 53 patients out of 1032 experienced

pelvic or abdominal infection in the intervention group compared

to 34 out of 1062 patients in the control group. These differences

did not reach statistical significance [OR: 1.92; 95% CI: 0.95,

3.92; p=0.07].

0.5% Ferric Hyaluranate gel: One study which used 0.5% ferric

hyaluranate gel reported occurrence of peritonitis in one patient

(of 17) in the intervention group while none were recorded in the

control group. However, there were no instances of intra-abdom-

inal or pelvic abscesses in the study.

[3] INCIDENCE OF ANASTOMOTIC LEAKS

The incidence of anastomotic leaks was reported in a total of 5

studies.

Hyaluranate/CMC membrane: Four studies using hyalu-

ranate/CMC membrane specifically reported on the incidence of

anastomotic leak (Becker 1996; Beck 2003; Tang 2003; Cohen

2005). The study by Tang reported two different phases in the

trial. In Phase I the intervention group had the ileostomy closed

at 3 weeks while the control group had the ileostomy closed at

6-12 weeks. As the radiologically evaluated anastomotic leak rate

is bound to be higher at 3 weeks when compared to 6-12 weeks,

potentially introducing a time bias between the two groups, pa-

tients from the phase I of the study were not included in the meta-

analysis of anastomotic leaks. However, in Phase II in the trial,

the intervention and control groups were treated similarly, with

ileostomy closure being performed at 3 weeks in both groups.

Data from Phase II of this trial was considered for meta-analysis.

In the study by Beck 2003, which was by far the largest study in

the group, there was a higher incidence of anastomotic leak and

leak related events (fistula, peritonitis, abdominopelvic abscess,

systemic sepsis) in the intervention group (76/882=8.6%) when

compared to the control group (46/909= 5.1%). In the subgroup

where HA/CMC membrane was wrapped around the anastomo-

sis, the incidence of these events was particularly high (39/289=

13.5%) and the difference was statistically significant when com-

pared to control in this particular study (Beck 2003).

When data from all three contributing studies was considered in

the meta-analysis, there was a total of 43 anastomotic leaks in 1066

patients in the treatment arm compared to 26 anastomotic leaks in

1098 patients in the control arm. The observed higher incidence

of anastomotic leak in the intervention group was not statistically

significant [OR: 1.61; 95%CI: 0.69, 3.71; p=0.27].

0.5% Ferric Hyaluranate gel: One study reported the incidence

of anastomotic leak when 0.5% ferric hyaluranate gel was com-

pared with control (Tang 2006). There were 5 anastomotic leaks

among 17 patients in the treatment group compared to 1 anas-

tomotic leak in 15 patients in the control group. The observed

fivefold increase in anastomotic leak in the treatment group was

however not statistically significant [OR: 5.83; 95% CI: 0.6, 57.1;

p=0.13].

[4] INCIDENCE OF ILEUS OR EARLY POSTOPERATIVE

OBSTRUCTION

Data on the incidence of ileus or early postoperative obstruction

not specifically attributed to adhesions was available in 5 studies.

Hyaluranate/CMC membrane: Four studies reported on the ef-

fect of use of this membrane agent on the incidence of ileus (Becker

1996; Beck 2003; Tang 2003; Cohen 2005). A total of 58 patients

out of 1113 developed ileus in the treatment arm compared to 57

out of 1148 patients in the control arm. These differences were not

statistically significant [OR: 1.05; 95% CI: 0.72, 1.53; p=0.79].

0.5% Ferric hyaluranate gel: The only study which reported on

the use of 0.5% Ferric hyaluranate gel found a nine fold higher

incidence of ileus in the intervention group (10 out of 17) com-

pared to the control group (2 out of 15) (Tang 2006). The dif-

ferences were statistically significant [OR: 9.29; 95% CI: 1.57,

54.77; p=0.01].

[5] INCIDENCE OF WOUND INFECTION

Incidence of wound infection was reported in a total of 5 studies.

Hyaluranate/CMC membrane: Four studies used Hyaluronic

acid/carboxymethyl cellulose membrane (Becker 1996; Vrijiland

2002; Beck 2003; Cohen 2005). A total of 49 wound infections

occurred in 1053 patients in the intervention group compared to

43 wound infections in 1083 patients in the control group. The

observed differences in wound infection between the intervention

and control groups were not statistically significant [OR 1.18;

95% CI: 0.77, 1.80; p=0.46].


surgery (Review)


0.5% Ferric hyaluranate gel: One study (Tang 2006) reported

wound infection rates when 0.5% Ferric hyaluronate gel was used

as the agent. This was an underpowered study as it had to be

prematurely terminated. There were 6 infections in 17 patients in

the intervention group compared to 3 infections in 15 patients

in the control group. There were no significant differences in the

incidence of wound infection between the two groups [OR: 2.18;

95% CI: 0.44, 10.9; p=0.34].

[6]INCIDENCE OF OVERALL MORTALITY

Overall mortality was reported in a total of 2 studies.

Hyaluranate/CMC membrane: In one study using Hyalu-

ranate/CMC membrane (Becker 1996), there was one mortality

among 91 patients in the treatment group (1%) and none among

92 patients in the control group. The differences were not statis-

tically significant.

0.5% Ferric hyaluranate gel: In one study using 0.5% ferric

hyaluranate gel (Tang 2006), there was one mortality each in

the treatment group (n=17; 5.8%) and the control group (n=15;

6.6%). These differences were not statistically significant.

D I S C U S S I O N

Peritoneal injury which occurs during laparotomy almost invari-

ably results in the formation of adhesions. It has been estimated

that up to a third of the adult population in the western world,

may be harbouring adhesions which carry a lifelong potential for

complications. The resultant burden to the patient, the surgeon

and the health care system on a global scale is high. From this per-

spective, research in peritoneal adhesions especially its prevention,

should be given a high priority. Methods to prevent adhesions may

be broadly grouped into surgical techniques and adjuncts. Sur-

gical techniques include such poorly defined suggestions such as

gentle peritoneal handling to more specific interventions such as

avoiding peritoneal closure, use of starch free gloves and avoiding

spillage of gall stones. However, these surgical techniques alone

are not completely effective on their own and hence the search

for adjuncts. Adjuncts may further be classified into physical (or

mechanical) types and chemical categories. The physical group

of agents are applied intraperitoneally and primarily act by form-

ing a barrier between adjacent peritoneal surfaces. They may be

membranes or sheets which have only a local effect where they

are applied (such as Hyaluronic acid/Carboxymethyl cellulose; ex-

panded poly-tetrafluoro ethylene; Oxidised regenerated cellulose)

or fluid agents which diffuse to coat a larger surface area of the peri-

toneum (such as 4% icodextrin solution and 0.5% ferric hyalu-

ranate gel). The chemical group of agents are more heterogenous

and have an effect by chemically interfering with adhesio genesis.

Examples include ani-inflammatory agents, anti-coagulants, anti-

fibrotic agents and proteolytic agents.

This review is limited to the evaluation of intra-peritoneally ap-

plied membrane and fluid agents. Over the last five decades, more

than a hundred different natural and synthetic agents have been

evaluated in animal experiments for their prophylactic potential.

While a large proportion of these pre-clinical studies have reported

beneficial effects, only a few of these agents have been evaluated in

clinical studies in humans. Most of the earlier studies in humans

were initially in the context of gynaecological operations and are

extensively reviewed elsewhere. It is only in the last decade that

clinical studies evaluating prophylactic agents in general surgical

operations started to emerge. To date, data on only two intra-

peritoneal agents namely hyaluronic acid/carboxymethyl cellulose

membrane and 0.5% ferric hyaluronate gel is available, in the con-

text of randomised controlled studies in general surgical opera-

tions. We found only a total of eight published randomised clin-

ical trials of prophylactic agents, seven of which used hyaluronic

acid/carboxymethyl cellulose membrane and one which reported

on 0.5% ferric hyaluranate gel. The probable reason for the scarcity

of published randomised trials in this field is, because of the fact

that, assessing the efficacy of adhesion prevention involves a sec-

ond invasive procedure such as laparoscopy or laparotomy, while

assessing some of the outcomes such as the effect on intestinal

obstruction needs long term follow up of the order of many years.

This also probably accounts for the generally small numbers of

patients recruited in individual studies as observed in the current

review, in which, despite the fact that five of the seven included

studies were multicentre studies, involving up to 22 centres, bar-

ring one study which recruited more than a thousand patients, the

total patients recruited in the individual studies was low resulting

in a median of 96 patients among the included studies. Absence

of blinding in a trial is known to be a source of bias and causes

overestimation of effects. Only four out of seven studies described

blinded outcome assessment. Though blinding may be difficult to

achieve in surgical patients because of the logistics involved and

is likely to be more expensive to implement, it would be a worth-

while means to achieve more accurate results. None of the studies

satisfied all the criteria for good quality studies as stated in the pro-

tocol. Five out of the seven studies were of intermediate quality,

while the other two were considered poor quality mainly due to

the high attrition rates in these studies. Trials in future, should aim

to clearly state the methods used to randomise, to conceal alloca-

tion and to achieve blinding in more detail. Similarly, they should

provide details of sample size calculations, should be adequately

powered to answer the key questions being addressed by the study.

Calculation of sample size and power should take into account

the attrition of patients during the course of the study, which was

unacceptably high in three of the seven studies included in the

review. The lack of explicit definitions in the outcome in some

studies and wide variations in the methods used to measure some

of the outcomes (such as severity of adhesions) made comparisons

amongst the studies difficult.

In a randomised clinical study, it is important that the interven-


surgery (Review)


tion and the control group are otherwise treated identically. Most

studies achieved this at the point of intervention and assessment.

However, Tang 2003, had an initial phase in the study (phase I),

where the intervention group were assessed early at 3 weeks while

the control group were assessed between 6-12 weeks. Such dif-

ferential treatment may potentially introduce bias and hence this

phase of the study has been discounted when analysing outcomes

such as leak rates. Variable follow up is another element which

may potentially affect reporting of outcomes. As discussed in the

section on methodological quality, there was a wide variation in

follow up and in some studies there was lack of clarity as to the

duration of follow up. There is a need for consensus among the

researchers as to the minimum acceptable follow up period for

various outcomes and future studies should aim to achieve this

for all patients available for analysis in the intervention and the

control groups.

When comparing the effect of membrane agents on leak rates, it

may be more relevant to compare the rates between patients in

whom the membrane was in contact with the anastomosis with

those in whom it was not (the control group). However, this was

not achieved in the studies that reported on leak rates. Becker 1996

and Cohen 2005 used the membrane only in contact with the

midline wound, while Tang 2003 used the membrane around the

stoma. In Beck 2003, there were patients in the intervention group

who did not have the membrane wrapped around the anastomosis.

The leak rate for this subgroup was similar to the control group.

But the leak rate was significantly higher for the subgroup which

had the membrane wrapped around the anastomosis. Because the

randomisation was not stratified by placement of the membrane,

the differences could not be explored further. The leak rates were

only compared between the intervention group as a whole and the

control group and these were not significantly different.

The results of this review should be interpreted in the light of

the shortcomings discussed above. In summary, from the avail-

able evidence, the use of Hyaluronic acid/Carboxymethyl cellu-

lose membrane reduced the incidence of adhesions [OR 0.15

(95% CI: 0.05, 0.43); p=0.0005] and was associated with a re-

duced extent of adhesions [WMD -25.9%(95%CI: -40.56, -

11.26); p=0.0005] and severity of adhesions. However, the use

of Hyaluronic acid/Carboxymethyl cellulose membrane did not

reduce the incidence of intestinal obstruction [odds ratio: 0.84

(95% CI: 0.34, 2.08)]. With respect to the incidence of adhe-

sive intestinal obstruction (needing surgical intervention), there

was no significant difference between the intervention and control

groups [Odds ratio: 0.84 (95% CI: 0.24, 2.7)].

On the basis of currently available evidence, Hyaluronic

acid/Carboxymethyl cellulose membrane has an acceptable safety

profile.The overall morbidity and mortality were comparable be-

tween the hyaluronic acid/carboxymethyl cellulose membrane

groups and the control groups as was the specific morbidity such

as wound infection, intra-abdominal sepsis, anastomotic leak and

prolonged ileus. There are concerns regarding the higher incidence

of anastomotic leak in a subgroup of patients where the membrane

is directly wrapped around the anastomosis (Beck 2003). How-

ever, this has not been specifically studied in a randomised trial.

No valid conclusions could be drawn on the efficacy of 0.5%

ferric hyaluranate gel, the only other agent on which data relevant

to the subject of this review is available. While there were no

differences in the mortality rates between the intervention and

control groups, there was a high incidence of overall morbidity in

the intervention group [OR 5.04; 95%CI:1.1; 22.9]. While the

rates of wound infection, intra-abdominal abscess and anastomotic

leak were comparable, there was a higher incidence of ileus in the

intervention group [OR: 9.29; 95% CI: 1.57, 54.77; p=0.01] .

A U T H O R S ’ C O N C L U S I O N S

Implications for practice

Based on the currently available data considered in this systematic

review, there is evidence that the use of HA/CMC membrane,

reduces the incidence, extent and severity of adhesions. This may

have implications in re-operative abdominal surgery in terms of

time, ease of access, blood loss, visceral injury and cost, but these

outcomes were beyond the remit of this review. With respect to

symptomatic long term outcomes, however, there is currently no

evidence that use of HA/CMC membrane, reduces the incidence

of intestinal obstruction or the need for operative intervention

in adhesive intestinal obstruction in the first few years following

the index abdominal surgery. HA/CMC membrane appears to be

safe, though there are concerns that it may increase the risk of

anastomotic leak when wrapped around the anastomosis. Hence,

in the absence of an ideal or alternative agent of proven efficacy

in general surgical patients, HA/CMC may be considered in the

prophylaxis of intra-peritoneal adhesions, at the discretion of the

surgeon, tailored to the individual clinical context.

0.5% Ferric hyaluranate gel, has an unacceptably high morbidity,

based on the limited data available, and hence cannot be recom-

mended.

Implications for research

a. Evidence on effectiveness of other agents in general surgery

Currently Hyaluronic acid/Carboxymethylcellulose (HA/CMC)

membrane remains the only agent on which evidence of effective-

ness as an anti-adhesio genesic agent is available in general surgical

patients. However, HA/CMC has certain shortcomings. It is suit-

able only for open surgery and it needs to be applied with some

degree of precision on to the raw surfaces to be effective. It does

not prevent adhesions in regions of the peritoneum which may

harbour areas of injury not be macroscopically visible, areas which

therefore, are not protected. Fluid agents such as 4% icodextrin


surgery (Review)


circumvent many of these problems and there is evidence from at

least one study in gynaecological surgery of its efficacy in prevent-

ing adhesions (diZerega 2002). Studies in general surgical opera-

tions are ongoing and would perhaps clarify its role in future. A

plethora of other pharmacological and physical (membrane and

fluid) agents have been shown to be effective in pre-clinical animal

experiments, but only a selected few have been tested in human

studies. Even fewer agents have been tried in randomised clini-

cal trials, mostly in gynaecological surgery. Adequately powered,

prospective, blinded, randomised controlled clinical trials in gen-

eral surgical patients using agents which have proven efficacy in

preclinical studies, may potentially uncover useful therapeutic op-

tions.

b. Evidence of synergism

The potential for synergistic effects of a combination of different

types of adhesion prevention strategies, such as anti-inflammatory

strategies, fibrinolytic strategies and barrier agents applied together

has been demonstrated in animal experiments. The clinical value

of using a combination of strategies in general surgical patients

undergoing laparotomy may be worth investigating.

c. Evidence on other clinically important effects

Though it is evident that film barrier agents reduce the incidence

and severity of adhesions, currently there is lack of evidence that

this translates to a reduction in symptomatic long term outcomes

such as incidence of intestinal obstruction and the proportion

needing surgical intervention for adhesive obstruction. If there

was indeed a small benefit from the barrier agents, which was

missed in the available studies (type II error), then this will need a

large, adequately powered study to uncover the effect. As adhesive

disease is common, even small risk reductions have the potential to

translate to significant benefits when applied on a global scale.The

other clinically significant long term outcome is chronic pain,

on which there is no data available at present on the efficacy of

prophylactic agents in general surgical patients.

d. Evidence of effectiveness in a range of operations

Generalised peritonitis results in a higher incidence of adhesion

related re-admission and re-operation. Data on safety and efficacy

of anti-adhesio genetic agents used in the context of pathologies

leading to generalised peritonitis is worth pursuing. Similarly use

of prophylactic agents in other types of surgery such as small bowel

resections for Crohns disease and a range of laparoscopic surgical

procedures would be valuable. If noninvasive tools for evaluation

of adhesions such as magnetic resonance imaging are validated in

assessing adhesions then a broader range of surgical interventions

and consequently a larger subgroup of patients could be studied.

e. Identification of high risk groups

It is known that adhesions and adhesion related complications are

more frequent after certain types of surgery such as pelvic surgery.

However, at present there is no validated system to identify indi-

vidual patients with an increased risk of adhesions and related com-

plications during the perioperative period. Future studies should

focus on collecting good quality, prospective data (epidemiolog-

ical, clinical, biochemical, cellular and biomolecular) which may

then be used to predict the risk in individual patients. Risk pre-

diction may help in prognostication, planning service needs and

targeting anti-adhesive therapy.

f. Cost effectiveness

Economic analysis in the field of peritoneal adhesion prevention

is understandably difficult for many reasons including controlling

for the number of potential confounding factors in the periop-

erative period, the difficulty in measuring the economic impact

accurately and the length of follow up required. Nevertheless as

economic analysis may be useful in resource planning it is perhaps

worthwhile exploring the cost effectiveness in a carefully chosen

index procedure which is amenable to standardisation.

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

None

R E F E R E N C E S

References to studies included in this review

Beck 2003 {published data only}

Beck DE, Cohen Z, Fleshman JW, Kaufman H, van Goor H, Wolff

BG for the adhesion study group steering committee. A prospective,

randomised, multicenter, controlled study of the safety of seprafilm

adhesion barrier in abdominopelvic surgery of the intestine. Dis

Colon Rectum 2003;46:1310–19.

Becker 1996 {published data only}

Becker JM, Dayton MT, Fazio VW, Beck DE, Stryker SJ, Wexner

SD, et al.Prevention of postoperative abdominal adhesions by a

sodium hyaluronate based bioresorbable membrane: a prospective,

randomised, double-blind multicenter study. J Am Coll Surg 1996;

183(4):297–306.

Cohen 2005 {published data only}

Cohen Z, Senagore A, Dayton M, Koruda M, Beck DE, Wolff BG,

et al.Prevention of postoperative abdominal adhesions by a novel,

Glycerol/sodium hyaluronate/ carboxymethylcellulose- based biore-

sorbable membrane: a prospective, randomised, evaluator blinded

multicenter study. Dis Colon Rectum 2005;48:1130–39.

Fazio 2006 {published data only}

Fazio VW, Cohen Z, Fleshman JW, van Goor H, Bauer JJ, Wolff

BG, et al.Reduction in adhesive small-bowel obstruction by seprafilm


surgery (Review)


ahesion barrier after intestinal resection. Dis Colon Rectum 2006;49

(1):1–11.

Kusunoki 2005 {published data only}

Kusunoki M, Ikeuchi H, Yanagi H, Noda M, Tonouchi H, Mohri Y,

et al.Bioresorbable hyaluronate-carboxymethylcellulose membrane

(Seprafilm) in surgery for rectal carcinoma: a prospective randomised

clinical trial. Surgery today 2005;35(11):940–5.

Tang 2003 {published data only}

Tang CL, Seow-Choen F, Fook-Chong S, Eu KW. Bioresorbable ad-

hesion barrier facilitates early closure of the defunctioning ileostomy

after rectal excision: a prospective, randomised trial.. Dis Colon Rec-

tum 2003;46(9):1200–7.

Tang 2006 {published data only}

Tang CL, Jayne DG, Seow-Choen F, Ng Y-Y, Eu K-W, Mustapha N.

A randomised controlled trial of 0.5% ferric hyaluronate gel (Intergel)

in the prevention of adhesions following abdominal surgery. Annals

of Surgery 2006;243:449–455.

Vrijiland 2002 {published data only}

Vrijiland WW, Tseng LNL, Eijkman HJM, Hop WCJ, Jakimowicz JJ,

Leguit P, et al.Fewer intraperitoneal adhesions with use of hyaluronic

acid-carboxymethylcellulose membrane: a randomised clinical trial.

Ann Surg 2002;235(2):193–99.

References to studies excluded from this review

Arnbjornsson 1983 {published data only}

Arnbjornsson E, Goransson G. The effect of intraperitoneal fluid on

the prevention of small intestinal obstruction after appendicectomy.

Ann Chir Gynaecol 1983;72:250–52.

Beck 1997 {published data only}

Beck DE. The role of seprafilm bioresorbable membrane in adhesion

prevention. Eur J Surg 1997;Suppl 577:49–55.

Kudo 2004 {published data only}

Kudo FA, Nishibe T, Miyazaki K, Murashita T, Nishibe M. Use of

bioresorbable membrane to prevent postoperative small bowel ob-

struction in transabdominal aortic aneurysm surgery. Surgery Today

2004;34:448–51..

Salum 2001 {published data only}

Salum MR, Lam DT, Wexner SD, Pikarsky A, Baig MK, Weiss EG,

et al.Does limited placement of bioresorbable membrane of modified

sodium hyaluronate and carboxymethyl cellulose (Seprafilm) have

possible short term beneficial effects?. Dis Colon Rectum 2001;44(5):

706–12.

Wang 1994 {published data only}

Wang RQ, Chen JZ, Ren GY. Intraperitoneal perfusion of compound

injection of salvia miltiorrhiza with dachengqi decoction in treating

adhesive intestinal obstruction. Zhongguo Zhong Xi Yi Jie He Za Zhi

1994;14(10):595–97.

Additional references

Ahmad 2008

Ahmad G, Duffy JM, Farquhar C, Vandekerckhove P, Watson A,

Wiseman D. Barrier agents for adhesion prevention after gynaeco-

logical surgery. Cochrane Database of Systematic Reviews 2008, Issue

2:. [DOI: 10.1002/14651858.CD000475.pub2]

Chalmers 1981

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

B, Reitman D, Ambroz A. A method for assessing the quality of

a randomised controlled trial. Control Clin Trials 1981;2:31–49.

[MEDLINE: 7261638]

Coleman 2000

Coleman MG, McLain AD, Moran BJ. Impact of previous surgery on

time taken for incision and division of adhesions during laparotomy.

Dis Colon Rectum 2000;43:1297–9. [MEDLINE: 11005501]

Cox 1993

Cox MR, Gunn IF, Eastman MC, Hunt RF, Heinz AW. The operative

aetiology and types of adhesions causing small bowel obstruction.

Aust N Z J Surg 1993;63(11):848–52. [MEDLINE: 8216061]

diZerega 1994

diZerega GS. Contemporary adhesion prevention. Fertil Steril 1994;

61:219–35. [MEDLINE: 8299773]

diZerega 2002

diZerega GS, Verco SJS, Young P, Kettle M, Koback W, Martin D,

Senfilippo J, Peers EM, Scrimgeour A, Brown CB. A randomised

controlled pilot study of the safety and efficacy of 4% icodextrin so-

lution in the reduction of adhesions following laparoscopic gynaeco-

logical surgery. Human Reproduction 2002;17:1031–1038.

Ellis 1997

Ellis H. The clinical significance of adhesions: Focus on intesti-

nal obstruction.. Eur J Surg 1997;577 (suppl):5–9. [MEDLINE:

9076446]

Fazio 1995

Fazio VW, Ziv Y, Church JM, Oakley JR, Lavery IC, Milsom JW,

Schroeder TK. Ileal pouch-anal anastamoses complications and func-

tion in 1005 patients. Ann Surg 1995;222:120–7. [MEDLINE:

7639579]

Jeekel 1997

Jeekel H. Cost implications of adhesions as highlighted in a European

study. Eur J Surg Suppl 1997;S79:43–5.

Lockhart-Mummery1967

Lockhart Mummery HE. Intestinal polyposis: the present position.

Proc Roy Soc Med 1967;60:381–4. [MEDLINE: 6021668]

Menzies 1990

Menzies D, Ellis H. Intestinal obstruction from adhesions- how big

is the problem?. Ann R Coll Surg Engl 1990;72:60–3. [MEDLINE:

2301905]

Menzies 1992

Menzies D. Peritoneal adhesions. Incidence, cause and prevention.

Ann Surg 1992 1992;24(1):29–45. [MEDLINE: 1727325]

Menzies 1993

Menzies D. Postoperative adhesions: their treatment and relevance

in clinical practice. Ann R Coll Surg Engl 1993;75:147–53. [MED-

LINE: 8323205]

Menzies 2001

Menzies D, Parker M, Hoare R, Knight A. Small bowel obstruction

due to postoperative adhesions: treatment patterns and associated

costs in 110 hospital admissions. Ann R Coll Surg Engl 2001;83:

40–6. [MEDLINE: 11212449]


surgery (Review)


Metwally 2006

Metwally M, Watson A, Lilford R, Vandekerckhove P. Fluid and

pharmacological agents for adhesion prevention after gynaecological

surgery. Cochrane Database of Systematic Reviews 2006 Apr 19, Issue

2:. [DOI: CD001298]

Nieuwenhuijzen 1998

Nieuwenhuijzen M, Reijnen MM, Kuijpers JH, van Goor H. Small

bowel obstuction after total or subtotal colectomy: a 10 year retro-

spective review. Br J Surg 1998;85:1242–5. [MEDLINE: 9752868]

Parker 2001

Parker MC, Ellis H, Moran BJ, Thompson JN, Wilson MS, Menzies

D, McGuire A, Lower AM, Hawthorn RJ, O’Briena F, Buchan S,

Crowe AM. Postoperative adhesions: Ten-year follow-up of 12,584

patients undergoing lower abdominal surgery. Dis Colon Rectum

2001;44:822–30. [MEDLINE: 11391142]

Ray 1998

Ray NF, Denton WG, Thamer M, Henderson SC, Perry S. Abdom-

inal adhesiolysis: inpatient care and expenditure in the united states

in 1994. J Am Coll Surg 1998;186(1):1–9. [MEDLINE: 9449594]

Risberg 1997

Risberg B. Adhesions: Preventive strategies. Eur J Surg 1997;577:

32–9. [MEDLINE: 9076450]

Van der Krabben 2000

Van Der Krabben AA, Dijkstra FR, Nieuwenhuijzen M, Reijnen

MM, Schaapveld M, Van Goor H. Morbidity and mortality of inad-

vertent enterotomy during adhesiotomy. Br J Surg 2000;87:467–71.

[MEDLINE: 10759744]

Wilkins 1986

Wilkins BM, Spitz L. Incidence of post-operative adhesive obstruc-

tion following neonatal laparotomy. Br J Surg 1986;73:762–4.

[MEDLINE: 3756446]

Wilson 1998

Wilson MS, Hawkswell JA, McCloy RF. The natural history of adhe-

sional small bowel obstruction: Counting the cost. Br J Surg 1998;

85:1294–8. [MEDLINE: 9752881]∗ Indicates the major publication for the study


surgery (Review)


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]

Beck 2003

Methods Prospective, randomised, multicenter , single-blind study.

Participants Patients undergoing abdominopelvic surgery

Total: 1791

Barrier: 882

Control: 909

Interventions Barrier group: received 3 to 10 pieces of Hyaluronic acid/ Carboxymethyl cellulose membrane to surgically

traumatised surfaces

Control group: No specific anti-adhesion prophylaxis.

Outcomes Complications occurring up to 1 month post-operatively were assessed.

OUTCOMES: Overall adverse events; Specific adverse events: Wound infection; anastomotic leak; fistula;

intra-abdominal abscess; sepsis; Pulmonary embolism

Notes Centralised randomisation using random number tables to generate randomisation sequence. Allocation

concealment not mentioned.

Risk of bias

Item Authors’ judgement Description

Allocation concealment? Unclear B - Unclear

Becker 1996

Methods Prospective, randomised, multicenter (11 centres), double-blind study.

Participants Patients undergoing colectomy and ileo-anal pouch procedure for ulcerative colitis or familial adenomatous

polyposis.

Total: 183

Barrier:

Control:

Evaluable: 175

Barrier: 85

Control: 90

Interventions Barrier group: received Hyaluronic acid/Carboxymethylcellulose membrane applied under the wound

and over the small bowel and omentum.

Control group: No specific anti-adhesion prophylaxis


surgery (Review)


Becker 1996 (Continued)

Outcomes Assessed laparoscopically by an allocation blinded observer, at second surgery done 8-12 weeks later for

ileostomy closure.

OUTCOMES: Incidence of adhesions; extent of adhesions; severity of adhesions; adverse effects.

Notes Randomisation was stratified and performed at the point of intervention. Method used to generate random

sequence was not specified.

Risk of bias



Cohen 2005

Methods Prospective, randomised, multicenter (12 centres), evaluator blinded study.

Intention to treat analysis.

Participants Patients undergoing colectomy and ileo-anal pouch procedure for ulcerative colitis or familial adeno-

matous polyposis.

Total: 120

Barrier: 59

Contreol: 61

Interventions Barrier group: received 2 to 4, Glycerol-Hyaluronic acid/Carboxymethyl cellulose membranes applied

under the wound and over the viscera and omentum.

Control group: No specific anti-adhesion prophylaxis

Outcomes Assessed laparoscopically by an allocation blinded observer, at second surgery done 8-12 weeks later for

ileostomy closure.

OUTCOMES: Incidence of adhesions; extent of adhesions; severity of adhesions; adverse effects.

Notes Method of generation of random sequence not mentioned. Randomisation however, was at the point

of intervention.

Risk of bias



Free of selective reporting? Yes


surgery (Review)


Fazio 2006

Methods prospective randomised, multi centre (22 centres), single blind study

Participants Patients undergoing bowel resection (large bowel, small bowel , rectum)

Total: 1701

Barrier: 840

Control: 861

Interventions Barrier group: received 3 to 10 pieces of Hyaluronic acid/ Carboxymethyl cellulose membrane to

surgically traumatised surfaces


Outcomes Assessed over a follow up ranging from 2 to 5 years (mean 3.4 years)

OUTCOMES: Incidence of bowel obstruction; time to first obstruction

Notes Randomisation by random number tables: appropriate . Randomisation revealed at the point of inter-

vention.

Risk of bias



Free of selective reporting? No There was a subgroup of 90 patients who were excluded from the main cohort

at the analysis stage (i.e post-randomisation). Excluding this group skewed

the results in favour of the intervention group (as presented in Fazio 2006),

where including the subgroup would have resulted in there being no statistically

significant difference between the intervention and control groups.

Kusunoki 2005

Methods Prospective randomised single centre, study

Participants Patients with rectal cancer (Stage II or greater) who underwent pre-operative radiotherapy, followed by

two staged surgery and chemotherapy.

Total: 62

Barrier:32

Control:30

Evaluable: 59

Barrier:30

Control: 29

Interventions Barrier group: Hyaluronic acid/ Carboxymethylcellolose membrane was placed under the midline laparo-

tomy wound and around the two limbs of the ileostomy.



surgery (Review)


Kusunoki 2005 (Continued)

Outcomes Evaluation of adhesions at the time of ileostomy closure

OUTCOMES: Severity of adhesions;

Extent of adhesions

Notes Method of generation of random sequence not mentioned. Allocation concealment was by sealed en-

velopes.

Risk of bias


Allocation concealment? Yes A - Adequate

Tang 2003

Methods Prospective randomised single centre, study

Participants Patients undergoing an elective rectal resection who needed an ileostomy

Total: 70

Barrier: 34

Control: 36

Evaluable: 43

Barrier: 21

Control: 22

Interventions Barrier group: Hyaluronic acid/ Carboxymethylcellolose membrane was wrapped around the two limbs

of the ileostomy.


Outcomes Evaluation of adhesions at ileostomy closure at 3 weeks.

OUTCOMES: adhesion scores assessing severity of adhesions ; stomal complications

Notes Method of generation of random sequence not mentioned. Allocation concealment was by sealed en-

velopes.

Risk of bias




surgery (Review)


Tang 2006

Methods Prospective randomised , single centre study

Participants Patients undergoing elective colorectal surgery

Total: 32

Barrier: 17

Control : 15

Evaluable: 32

Barrier: 17

Control: 15

Trial terminated prematurely due to high morbidity in the treatment group.

Interventions Barrier group: 300 mls of prewarmed, 0.5% Ferric Hyaluronate gel was instilled into the peritoneal cavity

at the end of the laparotomy procedure.

Control group: Prewarmed distilled water instilled into the peritoneal cavity.

Outcomes Evaluation by clinical follow up. No direct assessment of adhesions by planned laparoscopy or laparotomy

at a later date.

OUTCOMES: Morbidity, Mortality, incidence of intestinal obstruction. Need for subsequent adhesiol-

ysis, Quality of life.

Notes Computer generated random sequence. Alocation revealed only at the time of application of the inter-

vention by telephoning a centralised independent set up.

Risk of bias



Vrijiland 2002

Methods Prospective, randomised, multicenter study (8 centres)

Participants Patients undergoing Hartmanns resection for sigmoid diverticulitis or obstructed rectosigmoid.

Total: 71

Barrier: 32

Control: 35

Evaluable: 42

Barrier: 21

Control: 21

Interventions Barrier group: Received 1 to 3 Hyaluronic acid/Carboxymethylcellolose membranes applied to surgically

traumatised areas.



surgery (Review)


Vrijiland 2002 (Continued)

Outcomes Evaluation at second planned surgery. Blinded evaluation of adhesions.

OUTCOMES: Incidence of adhesions; severity of adhesions; complications

Notes Randomisation by computer generated random numbers: appropriate

Allocation concealment by sealed envelopes.

Risk of bias



Characteristics of excluded studies [ordered by study ID]

Arnbjornsson 1983 Non randomised study comparing 173 patients undergoing appendicectomy who received intra- peritoneal

normal saline, with 173 matched controls who received no prophylaxis.

Beck 1997 Dual publication of data. Data same as published in Becker JM 1996.

Kudo 2004 Non-randomised study comparing the incidence of early intestinal obstruction after transabdominal aortic

aneurysm surgery, between 21 patients who received Seprafilm and 30 patients who did not.

Salum 2001 Retrospective, non-randomised study comparing the incidence of intestinal obstruction between 259 patients

who underwent colorectal surgery (elective and emergency) in whom seprafilm was used, with 179 matched

controls.

Wang 1994 Non-randomised study comparing 47 patients who received an intra-peritoneal infusion of a Chinese herbal

compound injection containing Salvia miltiorrhiza with dachengqi decoction, with 38 matched controls who

received intraperitoneal antibiotics.


surgery (Review)


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

Comparison 1. HA/CMC membrane Versus Control

Outcome or subgroup titleNo. of

studies

No. of

participants Statistical method Effect size

1 Incidence of adhesions 3 335 Odds Ratio (M-H, Random, 95% CI) 0.15 [0.05, 0.43]

1.1 Overall 3 335 Odds Ratio (M-H, Random, 95% CI) 0.15 [0.05, 0.43]

2 Extent of adhesions 2 220 Mean Difference (IV, Random, 95% CI) -25.91 [-40.56, -

11.26]

2.1 Extent by percentage area 2 220 Mean Difference (IV, Random, 95% CI) -25.91 [-40.56, -

11.26]

3 Adverse events (Overall) 3 2094 Odds Ratio (M-H, Random, 95% CI) 1.19 [0.97, 1.46]

4 Wound infection 4 2136 Odds Ratio (M-H, Random, 95% CI) 1.18 [0.77, 1.80]

5 Intra-abdominal or pelvic

infection

3 2094 Odds Ratio (M-H, Random, 95% CI) 1.92 [0.94, 3.92]

6 Anastomotic leak 4 2164 Odds Ratio (M-H, Random, 95% CI) 1.61 [0.69, 3.71]

7 Fistula 1 1791 Odds Ratio (M-H, Random, 95% CI) 5.58 [1.62, 19.22]

8 Ileus 4 2261 Odds Ratio (M-H, Random, 95% CI) 1.05 [0.72, 1.53]

9 First episode of Intestinal

obstruction (All causes)


10 First episode of Intestinal

obstruction (Adhesive) needing

surgery


10.1 Original 2 1763 Odds Ratio (M-H, Random, 95% CI) 0.50 [0.27, 0.92]

10.2 Faz 90 0 0 Odds Ratio (M-H, Random, 95% CI) Not estimable

Comparison 2. Ferric Hyaluranate gel Versus Control

Outcome or subgroup titleNo. of

studies

No. of

participants Statistical method Effect size

1 Adverse events (Overall) 1 32 Odds Ratio (M-H, Random, 95% CI) 5.04 [1.11, 22.97]

2 Protracted Ileus 1 32 Odds Ratio (M-H, Random, 95% CI) 9.29 [1.57, 54.77]

3 Anastomotic leak 1 32 Odds Ratio (M-H, Fixed, 95% CI) 5.83 [0.60, 57.10]

4 Wound infection 1 32 Odds Ratio (M-H, Fixed, 95% CI) 2.18 [0.44, 10.91]

5 Mortality 1 32 Odds Ratio (M-H, Fixed, 95% CI) 0.88 [0.05, 15.33]


surgery (Review)


Analysis 1.1. Comparison 1 HA/CMC membrane Versus Control, Outcome 1 Incidence of adhesions.

Review: Intra-peritoneal prophylactic agents for preventing adhesions and adhesive intestinal obstruction after non-gynaecological abdominal surgery

Comparison: 1 HA/CMC membrane Versus Control

Outcome: 1 Incidence of adhesions

Study or subgroup Treatment Control Odds Ratio Weight Odds Ratio

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

1 Overall

Becker 1996 42/85 85/90 38.7 % 0.06 [ 0.02, 0.16 ]

Vrijiland 2002 16/21 19/21 22.8 % 0.34 [ 0.06, 1.98 ]

Cohen 2005 39/58 54/60 38.5 % 0.23 [ 0.08, 0.62 ]

Total (95% CI) 164 171 100.0 % 0.15 [ 0.05, 0.43 ]

Total events: 97 (Treatment), 158 (Control)

Heterogeneity: Tau2 = 0.54; Chi2 = 4.92, df = 2 (P = 0.09); I2 =59%

Test for overall effect: Z = 3.46 (P = 0.00054)

0.005 0.1 1 10 200

Favours treatment Favours control

Analysis 1.2. Comparison 1 HA/CMC membrane Versus Control, Outcome 2 Extent of adhesions.



Outcome: 2 Extent of adhesions

Study or subgroup Treatment Control Mean Difference Weight Mean Difference

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

1 Extent by percentage area

Becker 1996 42 48 (34) 85 67 (31) 53.9 % -19.00 [ -31.21, -6.79 ]

Cohen 2005 39 31 (37) 54 65 (34) 46.1 % -34.00 [ -48.73, -19.27 ]

Total (95% CI) 81 139 100.0 % -25.91 [ -40.56, -11.26 ]



-100 -50 0 50 100



surgery (Review)


Analysis 1.3. Comparison 1 HA/CMC membrane Versus Control, Outcome 3 Adverse events (Overall).



Outcome: 3 Adverse events (Overall)



Beck 2003 249/882 223/909 95.9 % 1.21 [ 0.98, 1.49 ]

Becker 1996 82/91 86/92 3.7 % 0.64 [ 0.22, 1.87 ]

Cohen 2005 59/59 60/61 0.4 % 2.95 [ 0.12, 73.88 ]

Total (95% CI) 1032 1062 100.0 % 1.19 [ 0.97, 1.46 ]


Heterogeneity: Tau2 = 0.0; Chi2 = 1.63, df = 2 (P = 0.44); I2 =0.0%


0.1 0.2 0.5 1 2 5 10


Analysis 1.4. Comparison 1 HA/CMC membrane Versus Control, Outcome 4 Wound infection.



Outcome: 4 Wound infection



Beck 2003 28/882 27/909 62.7 % 1.07 [ 0.63, 1.83 ]

Becker 1996 10/91 9/92 20.0 % 1.14 [ 0.44, 2.95 ]

Cohen 2005 5/59 4/61 9.7 % 1.32 [ 0.34, 5.17 ]

Vrijiland 2002 6/21 3/21 7.6 % 2.40 [ 0.51, 11.26 ]

Total (95% CI) 1053 1083 100.0 % 1.18 [ 0.77, 1.80 ]




0.1 0.2 0.5 1 2 5 10



surgery (Review)


Analysis 1.5. Comparison 1 HA/CMC membrane Versus Control, Outcome 5 Intra-abdominal or pelvic

infection.



Outcome: 5 Intra-abdominal or pelvic infection



Becker 1996 7/91 2/92 16.2 % 3.75 [ 0.76, 18.56 ]

Beck 2003 37/882 29/909 62.8 % 1.33 [ 0.81, 2.18 ]

Cohen 2005 9/59 3/61 21.0 % 3.48 [ 0.89, 13.56 ]

Total (95% CI) 1032 1062 100.0 % 1.92 [ 0.94, 3.92 ]




0.1 0.2 0.5 1 2 5 10


Analysis 1.6. Comparison 1 HA/CMC membrane Versus Control, Outcome 6 Anastomotic leak.



Outcome: 6 Anastomotic leak

Study or subgroup Treatment Control Odds Ratio Odds Ratio


Becker 1996 0/91 0/92 0.0 [ 0.0, 0.0 ]

Beck 2003 34/882 18/909 1.98 [ 1.11, 3.54 ]

Tang 2003 6/34 8/36 0.75 [ 0.23, 2.44 ]

Cohen 2005 3/59 0/61 7.62 [ 0.39, 150.78 ]

Total (95% CI) 1066 1098 1.61 [ 0.69, 3.71 ]




0.1 0.2 0.5 1 2 5 10



surgery (Review)


Analysis 1.7. Comparison 1 HA/CMC membrane Versus Control, Outcome 7 Fistula.



Outcome: 7 Fistula



Beck 2003 16/882 3/909 100.0 % 5.58 [ 1.62, 19.22 ]

Total (95% CI) 882 909 100.0 % 5.58 [ 1.62, 19.22 ]


Heterogeneity: not applicable


0.1 0.2 0.5 1 2 5 10


Analysis 1.8. Comparison 1 HA/CMC membrane Versus Control, Outcome 8 Ileus.



Outcome: 8 Ileus



Becker 1996 6/91 6/92 10.3 % 1.01 [ 0.31, 3.26 ]

Beck 2003 40/882 41/909 71.1 % 1.01 [ 0.64, 1.57 ]

Tang 2003 7/81 6/86 11.0 % 1.26 [ 0.41, 3.93 ]

Cohen 2005 5/59 4/61 7.6 % 1.32 [ 0.34, 5.17 ]

Total (95% CI) 1113 1148 100.0 % 1.05 [ 0.72, 1.53 ]




0.1 0.2 0.5 1 2 5 10



surgery (Review)


Analysis 1.9. Comparison 1 HA/CMC membrane Versus Control, Outcome 9 First episode of Intestinal

obstruction (All causes).



Outcome: 9 First episode of Intestinal obstruction (All causes)



Kusunoki 2005 2/30 5/29 20.7 % 0.34 [ 0.06, 1.93 ]

Fazio 2006 109/882 106/909 79.3 % 1.07 [ 0.80, 1.42 ]

Total (95% CI) 912 938 100.0 % 0.84 [ 0.34, 2.08 ]




0.1 0.2 0.5 1 2 5 10


Analysis 1.10. Comparison 1 HA/CMC membrane Versus Control, Outcome 10 First episode of Intestinal

obstruction (Adhesive) needing surgery.



Outcome: 10 First episode of Intestinal obstruction (Adhesive) needing surgery



1 Original

Fazio 2006 15/840 29/861 93.1 % 0.52 [ 0.28, 0.98 ]

Kusunoki 2005 1/32 3/30 6.9 % 0.29 [ 0.03, 2.96 ]

Subtotal (95% CI) 872 891 100.0 % 0.50 [ 0.27, 0.92 ]




2 Faz 90

Subtotal (95% CI) 0 0 0.0 % 0.0 [ 0.0, 0.0 ]



0.005 0.1 1 10 200


(Continued . . . )


surgery (Review)


(. . . Continued)Study or subgroup Treatment Control Odds Ratio Weight Odds Ratio


Test for overall effect: not applicable

Total (95% CI) 872 891 100.0 % 0.50 [ 0.27, 0.92 ]




0.005 0.1 1 10 200


Analysis 2.1. Comparison 2 Ferric Hyaluranate gel Versus Control, Outcome 1 Adverse events (Overall).


Comparison: 2 Ferric Hyaluranate gel Versus Control

Outcome: 1 Adverse events (Overall)



Tang 2006 11/17 4/15 100.0 % 5.04 [ 1.11, 22.97 ]

Total (95% CI) 17 15 100.0 % 5.04 [ 1.11, 22.97 ]




0.1 0.2 0.5 1 2 5 10



surgery (Review)


Analysis 2.2. Comparison 2 Ferric Hyaluranate gel Versus Control, Outcome 2 Protracted Ileus.



Outcome: 2 Protracted Ileus



Tang 2006 10/17 2/15 100.0 % 9.29 [ 1.57, 54.77 ]

Total (95% CI) 17 15 100.0 % 9.29 [ 1.57, 54.77 ]




0.1 0.2 0.5 1 2 5 10


Analysis 2.3. Comparison 2 Ferric Hyaluranate gel Versus Control, Outcome 3 Anastomotic leak.



Outcome: 3 Anastomotic leak


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

Tang 2006 5/17 1/15 100.0 % 5.83 [ 0.60, 57.10 ]

Total (95% CI) 17 15 100.0 % 5.83 [ 0.60, 57.10 ]




0.1 0.2 0.5 1 2 5 10



surgery (Review)


Analysis 2.4. Comparison 2 Ferric Hyaluranate gel Versus Control, Outcome 4 Wound infection.



Outcome: 4 Wound infection



Tang 2006 6/17 3/15 100.0 % 2.18 [ 0.44, 10.91 ]

Total (95% CI) 17 15 100.0 % 2.18 [ 0.44, 10.91 ]




0.1 0.2 0.5 1 2 5 10


Analysis 2.5. Comparison 2 Ferric Hyaluranate gel Versus Control, Outcome 5 Mortality.



Outcome: 5 Mortality



Tang 2006 1/17 1/15 100.0 % 0.88 [ 0.05, 15.33 ]

Total (95% CI) 17 15 100.0 % 0.88 [ 0.05, 15.33 ]




0.1 0.2 0.5 1 2 5 10


W H A T ’ S N E W

Last assessed as up-to-date: 2 August 2008.


surgery (Review)


23 October 2008 Amended Copy edited version for publication

3 August 2008 Amended Substantive amendment

H I S T O R Y

Protocol first published: Issue 1, 2005

Review first published: Issue 1, 2009

8 June 2004 New citation required and conclusions have changed Substantive amendment

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

SK: Development of protocol, literature search, contacting authors, publishers and other sources of information where necessary,

choosing relevant papers for review, critical appraisal of included and excluded studies, data extraction and analysis, writing the final

version of the review.

PFW: Literature search, critical appraisal of included and excluded studies, data extraction and analysis, participate in writing the final

version of the review.

DJL: Oversee the drafting of the protocol, Scrutinise papers and participate in decisions regarding inclusion and exclusion of studies,

supervise data extraction and analysis, participate and supervise the writing of the final version of he review.

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

There was no potential or actual conflict of interest perceived by the authors. This review did not receive any external funding.

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

In the protocol it was stated that, for inclusion in the review, the length of follow up for studies reporting on long term outcomes

should be at least 5 years. This was ideal. But among the available studies which satisfied all other criteria for inclusion, there were no

studies with follow up of 5 years of more. Hence, in the review, it was decided to include studies even with a follow up shorter than the

ideal of 5 years, as stated in the protocol.

Similarly, for studies reporting on short term outcomes, pertaining to safety, studies with a shorter follow up less than the ideal of 1

year, were also included in the review.


surgery (Review)


I N D E X T E R M SMedical Subject Headings (MeSH)

∗Membranes, Artificial; Abdomen [∗surgery]; Hyaluronic Acid [administration & dosage; adverse effects; ∗analogs & derivatives;∗therapeutic use]; Intestinal Obstruction [∗prevention & control]; Postoperative Complications [∗prevention & control]; Randomized

Controlled Trials as Topic; Tissue Adhesions [prevention & control]

MeSH check words

Humans


surgery (Review)


Intra-peritoneal prophylactic agents for preventing adhesions and adhesive intestinal obstruction after non-gynaecological abdominal surgery

Documents