Fluid and pharmacological agents for adhesion prevention after gynaecological surgery (Review)

Transcription

1 Fluid and pharmacological agents for adhesion prevention after gynaecological surgery (Review) Ahmad G, Mackie FL, Iles DA, O Flynn H, Dias S, Metwally M, Watson A This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2014, Issue 7

2 T A B L E O F C O N T E N T S HEADER ABSTRACT PLAIN LANGUAGE SUMMARY SUMMARY OF FINDINGS FOR THE MAIN COMPARISON BACKGROUND OBJECTIVES METHODS RESULTS Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure ADDITIONAL SUMMARY OF FINDINGS DISCUSSION AUTHORS CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES CHARACTERISTICS OF STUDIES DATA AND ANALYSES Analysis 1.1. Comparison 1 Hydroflotation agent vs no hydroflotation agent, Outcome 1 Improvement in pelvic pain at second-look laparoscopy Analysis 1.2. Comparison 1 Hydroflotation agent vs no hydroflotation agent, Outcome 2 Live birth rate Analysis 1.3. Comparison 1 Hydroflotation agent vs no hydroflotation agent, Outcome 3 Improvement in adhesion score at SLL Analysis 1.4. Comparison 1 Hydroflotation agent vs no hydroflotation agent, Outcome 4 Number of participants with worsening adhesion score Analysis 1.5. Comparison 1 Hydroflotation agent vs no hydroflotation agent, Outcome 5 Number of participants with adhesions at second-look laparoscopy Analysis 1.6. Comparison 1 Hydroflotation agent vs no hydroflotation agent, Outcome 6 Mean adhesion score at secondlook laparoscopy Analysis 1.7. Comparison 1 Hydroflotation agent vs no hydroflotation agent, Outcome 7 Clinical pregnancy rate.. 90 Analysis 1.9. Comparison 1 Hydroflotation agent vs no hydroflotation agent, Outcome 9 Ectopic pregnancy rate (per pregnancy) Analysis 2.3. Comparison 2 Gel agent vs no treatment, Outcome 3 Number of participants with improvement in adhesion score Analysis 2.4. Comparison 2 Gel agent vs no treatment, Outcome 4 Number of participants with worsening adhesion score Analysis 2.5. Comparison 2 Gel agent vs no treatment, Outcome 5 Number of participants with adhesions at second-look laparoscopy Analysis 2.6. Comparison 2 Gel agent vs no treatment, Outcome 6 Mean adhesion score at second-look laparoscopy. 93 Analysis 3.3. Comparison 3 Gel agent vs hydroflotation agent when used as an instillant, Outcome 3 Number of participants with improvement in adhesion score Analysis 3.4. Comparison 3 Gel agent vs hydroflotation agent when used as an instillant, Outcome 4 Number of participants with worsening adhesion score i

3 Analysis 3.5. Comparison 3 Gel agent vs hydroflotation agent when used as an instillant, Outcome 5 Number of participants with adhesions at second-look laparoscopy Analysis 3.6. Comparison 3 Gel agent vs hydroflotation agent when used as an instillant, Outcome 6 Mean adhesion score at second-look laparoscopy Analysis 4.2. Comparison 4 Steroid (any route) vs no steroid, Outcome 2 Live birth rate Analysis 4.3. Comparison 4 Steroid (any route) vs no steroid, Outcome 3 Number of participants with improvement in adhesion score Analysis 4.4. Comparison 4 Steroid (any route) vs no steroid, Outcome 4 Number of participants with worsening adhesion score Analysis 4.7. Comparison 4 Steroid (any route) vs no steroid, Outcome 7 Clinical pregnancy rate Analysis 4.9. Comparison 4 Steroid (any route) vs no steroid, Outcome 9 Ectopic pregnancy rate (per pregnancy).. 98 Analysis 5.4. Comparison 5 Intraperitoneal noxytioline vs no treatment, Outcome 4 Number of participants with worsening adhesion score Analysis 5.7. Comparison 5 Intraperitoneal noxytioline vs no treatment, Outcome 7 Clinical pregnancy rate Analysis 5.9. Comparison 5 Intraperitoneal noxytioline vs no treatment, Outcome 9 Ectopic pregnancy rate (per pregnancy) Analysis 6.3. Comparison 6 Intraperitoneal heparin solution vs no intraperitoneal heparin, Outcome 3 Number of participants with improvement in adhesion score Analysis 6.4. Comparison 6 Intraperitoneal heparin solution vs no intraperitoneal heparin, Outcome 4 Number of participants with worsening adhesion score Analysis 7.3. Comparison 7 Systemic promethazine vs no promethazine, Outcome 3 Number of participants with improvement in adhesion score Analysis 7.4. Comparison 7 Systemic promethazine vs no promethazine, Outcome 4 Number of participants with worsening adhesion score APPENDICES WHAT S NEW HISTORY CONTRIBUTIONS OF AUTHORS DECLARATIONS OF INTEREST SOURCES OF SUPPORT DIFFERENCES BETWEEN PROTOCOL AND REVIEW INDEX TERMS ii

4 [Intervention Review] Fluid and pharmacological agents for adhesion prevention after gynaecological surgery Gaity Ahmad 1, Fiona L Mackie 1, David A Iles 2, Helena O Flynn 3, Sofia Dias 4, Mostafa Metwally 5, Andrew Watson 6 1 Obstetrics & Gynaecology, Pennine Acute NHS Trust, Manchester, UK. 2 Obstetrics and Gynaecology, Blackpool Victoria Hospital, Blackpool, UK. 3 University Hospital of South Manchester, Manchester, UK. 4 School of Social and Community Medicine, University of Bristol, Bristol, UK. 5 The Jessop Wing and Royal Hallamshire Hospital, Sheffield Teaching Hospitals, Sheffield, UK. 6 Tameside & Glossop Acute Services NHS Trust, Tameside General Hospital, Ashton-Under-Lyne, UK Contact address: Gaity Ahmad, Obstetrics & Gynaecology, Pennine Acute NHS Trust, Manchester, UK. gaityahmad@hotmail.com. gaityahmad@aol.com. Editorial group: Cochrane Menstrual Disorders and Subfertility Group. Publication status and date: New search for studies and content updated (conclusions changed), published in Issue 7, Review content assessed as up-to-date: 7 April Citation: Ahmad G, Mackie FL, Iles DA, O Flynn H, Dias S, Metwally M, Watson A. Fluid and pharmacological agents for adhesion prevention after gynaecological surgery. Cochrane Database of Systematic Reviews 2014, Issue 7. Art. No.: CD DOI: / CD pub4. Background A B S T R A C T Adhesions are fibrin bands that are a common consequence of gynaecological surgery. They are caused by various conditions including pelvic inflammatory disease and endometriosis. Adhesions are associated with considerable co-morbidity, including pelvic pain, subfertility and small bowel obstruction. Patients may require further surgery-a fact that has financial implications. Objectives To evaluate the role of fluid and pharmacological agents used as adjuvants in preventing formation of adhesions after gynaecological surgery. Search methods The following databases were searched up to April 2014: Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, Cumulative Index to Nursing and Allied Health Literature (CINAHL) and PsycINFO. Studies involving hydroflotation, gel and such pharmacological agents as steroids, noxytioline, heparin, promethazine, N,O-carboxymethyl chitosan and gonadotrophinreleasing hormone agonists were evaluated. Selection criteria Randomised controlled trials investigating the use of fluid and pharmacological agents to prevent adhesions after gynaecological surgery. Gels were defined as fluid agents. Data collection and analysis Three review authors independently assessed trials for eligibility, extracted data and evaluated risk of bias. Results were expressed as odds ratios (ORs), mean differences (MDs) or standard mean differences (SMDs) as appropriate, with 95% confidence intervals (CIs). 1

5 Main results Twenty-nine trials were included (3227 participants), and nine were excluded. One study examined pelvic pain and found no evidence of a difference between use of hydroflotation agents and no treatment. We found no evidence that any of the antiadhesion agents significantly affected the live birth rate. When gels were compared with no treatment or with hydroflotation agents at second-look laparoscopy (SLL), fewer participants who received a gel showed a worsening adhesion score when compared with those who received no treatment (OR 0.16, 95% CI 0.04 to 0.57, P value 0.005, two studies, 58 women, I 2 = 0%, moderate-quality evidence) and with those given hydroflotation agents (OR 0.28, 95% CI 0.12 to 0.66, P value 0.003, two studies, 342 women, I 2 = 0%, high-quality evidence). Participants who received steroids were less likely to have a worsening adhesion score (OR 0.27, 95% CI 0.12 to 0.58, P value , two studies, 182 women, I 2 = 0%, low-quality evidence). Participants were less likely to have adhesions at SLL if they received a hydroflotation agent or gel than if they received no treatment (OR 0.34, 95% CI 0.22 to 0.55, P value < , four studies, 566 participants, I 2 = 0%, high-quality evidence; OR 0.25, 95% CI 0.11 to 0.56, P value , four studies, 134 women, I 2 = 0%, high-quality evidence, respectively). When gels were compared with hydroflotation agents, participants who received a gel were less likely to have adhesions at SLL than those who received a hydroflotation agent (OR 0.36, 95% CI 0.19 to 0.67, P value 0.001, two studies, 342 women, I 2 = 0%, high-quality evidence). No studies evaluated quality of life. In all studies apart from one, investigators stated that they were going to assess serious adverse outcomes associated with treatment agents, and no adverse effects were reported. Results suggest that for a woman with a 77% risk of developing adhesions without treatment, the risk of developing adhesions after use of a gel would be between 26% and 65%. For a woman with an 83% risk of worsening of adhesions after no treatment at initial surgery, the chance when a gel is used would be between 16% and 73%. Similarly, for hydroflotation fluids for a woman with an 84% chance of developing adhesions with no treatment, the risk of developing adhesions when hydroflotation fluid is used would be between 53% and 73%. Several of the included studies could not be included in a meta-analysis: The findings of these studies broadly agreed with the findings of the meta-analyses. The quality of the evidence, which was assessed using the GRADE approach, ranged from low to high. The main reasons for downgrading of evidence included imprecision (small sample sizes and wide confidence intervals) and poor reporting of study methods. Authors conclusions Gels and hydroflotation agents appear to be effective adhesion prevention agents for use during gynaecological surgery, but no evidence indicates that they improve fertility outcomes or pelvic pain, and further research is required in this area. Future studies should measure outcomes in a uniform manner, using the modified American Fertility Society (mafs) score. Statistical findings should be reported in full. P L A I N L A N G U A G E S U M M A R Y Use of fluids and pharmacological agents (medicinal drugs) to prevent the formation of adhesions (scar tissue) after surgery of the female pelvis Review question: This Cochrane systematic review evaluated all fluid and pharmacological agents that aim to prevent adhesion formation after gynaecological surgery (gels were defined as fluid agents). Background: Adhesions are defined as internal scar tissue that may form as part of the body s healing process after surgery. They can also be caused by pelvic infection and endometriosis. Adhesions join together tissues and organs that are not normally connected. They are common after gynaecological surgery and can cause pelvic pain, infertility and bowel obstruction. Women with adhesions may need further surgery, which is more difficult and can lead to additional complications. The fluid agents are placed inside the pelvic cavity (which contains all female reproductive organs) during surgery and physically prevent raw, healing tissues from touching. These fluids can be broken down into hydroflotation agents or gels; hydroflotation agents are fluids placed in large volumes (usually around a litre); gels are directly applied to the internal surgical site. Pharmacological agents act by changing part of the healing process. Study characteristics: We included 29 randomised controlled trials in the review (3227 participants). Of these, results of 18 trials were pooled (2740 participants). Results from the remaining 11 trials could not be used in the meta-analysis because investigators did not use a way of measuring adhesions that would allow findings to be pooled with other data, or because important statistical information was not reported. We searched all evidence up to April

6 Key results: Only one study evaluated pelvic pain and provided no evidence that the adhesion prevention agent made a difference. No evidence suggests that any of the investigated agents affected live birth rate. Regarding adhesions, participants given a fluid agent during surgery were less likely to form adhesions than participants who did not receive a fluid agent. When fluids and gels were compared with each other, gels appeared to perform better than fluids. No pharmacological agents showed good evidence of causing a significant effect on adhesions. No studies looked at differences in quality of life. All studies apart from one stated that investigators were going to assess serious adverse outcomes associated with the agents, and no adverse effects were reported. For gels, results suggest that for a woman with a 77% risk of developing adhesions without treatment, the risk of developing adhesions after a gel is used would be between 26% and 65%. For a woman with an 83% risk of worsening of adhesions after no treatment at initial surgery, the chance when a gel is used would be between 16% and 73%. Similarly, for hydroflotation fluids in a woman with an 84% chance of developing adhesions with no treatment, the risk of developing adhesions when hydroflotation fluid is used would be between 53% and 73%. Fluids and gels appear to be effective in reducing adhesions, but more information is needed to determine whether this affects pelvic pain, live birth rate, quality of life and long-term complications such as bowel obstruction. Further large, high-quality studies should be conducted in which investigators use the standard way of measuring adhesions as developed by the American Fertility Society (the modified AFS score). Quality of the evidence: The quality of the evidence ranged from low to high. The main reasons for downgrading of evidence were imprecision (small sample sizes and wide confidence intervals) and poor reporting of study methods. 3

7 S U M M A R Y O F F I N D I N G S F O R T H E M A I N C O M P A R I S O N [Explanation] Hydroflotation agents vs no hydroflotation agents for adhesion prevention after gynaecological surgery Patient or population: women after gynaecological surgery Settings: postsurgical Intervention: hydroflotation agents vs no hydroflotation agents Outcomes Illustrative comparative risks*(95% CI) Relative effect (95% CI) Improvement in pelvic pain in participants with a primary diagnosis of pelvic pain, at secondlook laparoscopy Assumed risk Corresponding risk No hydroflotation agents Hydroflotation agents 806per per1000 ( ) Live birth rate 140 per per 1000 (45-205) Improvement in adhesion score 437per per1000 ( ) OR 0.65 ( ) OR 0.67 ( ) OR 1.27 ( ) No. of participants (studies) 286 (1 study) 208 (2 studies) 665 (4 studies) Quality of the evidence (GRADE) moderate 1 moderate 1 moderate 1,2 Comments Number of participants with worsening adhesion score 308per per1000 (30-350) OR 0.28 ( ) 53 (1 study) moderate 1,3 Number of participants with adhesions at second-look laparoscopy 836per per1000 ( ) OR 0.34 ( ) 566 (4 studies) high 4

8 Mean adhesion score at second-look laparoscopy The mean adhesion score at second-look laparoscopy in the intervention groups was 0.06 standard deviations lower (0.2 lower-0.09 higher) Clinical pregnancy rate 234 per per 1000 (99-258) OR 0.64 ( ) 722 (4 studies) 310 (3 studies) high moderate 1 SMD-0.06(-0.2to0.09) 4 *The basis for the assumed risk is the median control group risk across studies. The corresponding risk(and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention(and its 95% CI). CI: Confidence interval; OR: Odds ratio. GRADE Working Group grades of evidence. High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Lowquality:Furtherresearchisverylikelytohaveanimportantimpactonourconfidenceintheestimateofeffectandislikelytochangetheestimate. Very low quality: We are very uncertain about the estimate. 1 Wide95%CI. 2 Moderateheterogeneity. 3 Smallnumberofevents. 4 Scale: mean of the mean adhesion score used. A lower mean mean adhesion score represents an improvement in adhesion disease. A variety of adhesion scoring systems were used (e.g. Hulka, mafs or system developed by authors for purpose of study); therefore for comparison, standardised mean difference was calculated. 5

9 B A C K G R O U N D Description of the condition Adhesions are fibrin bands that form as the result of aberrant peritoneal healing (Cheong 2001). Normally, peritoneal damage causes an inflammatory response; this activates the coagulation cascade, and a fibrin plug is formed over the damaged mesothelium, which is then broken down to reveal regenerated peritoneum. However, in adhesion formation, fibrinolysis of the fibrin plug is decreased and, consequently, a fibrin matrix develops. Adhesions may be defined as de novo, meaning that they have formed at a location that was previously free from adhesions, or re-formed, which describes adhesions that recur post adhesiolysis. A variety of factors influence the extent of adhesion formation, including type of surgery performed (i.e. laparoscopic or open), haemostasis and the presence of endometriosis and infection, particularly pelvic inflammatory disease (Diamond 2001). Although the aetiologies are different, the basic pathogenesis is similar. Description of the intervention Adhesion prevention agents can be divided into three types: fluid, pharmacological and barrier. This review will examine fluid and pharmacological agents. A separate review evaluates barrier agents. Fluid agents include both hydroflotation products and gels. Examples of hydroflotation devices are 4% icodextrin solution (Adept, Baxter, Berkshire, UK), an iso-osmolar and non-viscous high molecular weight glucose polymer, and 32% dextran (Hyskon Pharmacia, Uppsala, Sweden), a polysaccharide-containing solution that is no longer approved for use as an antiadhesion agent. Both agents can be used as intraperitoneal irrigants and/or instillants. Derivatives of hyaluronic acid form the basis of a number of antiadhesion gels. Hyaluronic acid is a linear polysaccharide with repeating disaccharide units composed of sodium D-glucuronate and N- acetyl-d-glucosamine. SepraSpray (Genzyme Corporation, Cambridge, MA, USA) contains hyaluronic acid in addition to carboxymethylcellulose powder and is applied to relevant tissues with the use of a preloaded delivery device. SepraCoat (Genzyme Corporation) is a dilute hyaluronic acid solution that is applied before and after surgery. Hyalobarrier gel (Nordic Pharma, Reading, UK) contains auto-cross-linked hyaluronic acid. Intergel (Gynecare, Lifecore Biomedical, Chaska, MN, USA) contains ferrous hyaluronic acid, although it has been withdrawn from the market. N,O-carboxymethyl chitosan is a derivative of chitin and is similar in structure to hyaluronic acid and carboxymethylcellulose. It is formed when the gel and solution components are combined. Polyethylene glycol (PEG)-based gels are also available. CoSeal (Baxter) is formed by mixing a powder and a liquid intraoperatively, both of which contain PEG and are then applied as a gel to relevant surfaces using a specific instrument. SprayGel (Confluent Surgical Inc., Waltham, MA, USA) is formed by two PEGcontaining liquid precursors, which create a cross-linked gel when combined. Intercoat (FzioMed, San Luis Obispo, CA, USA) is an Oxiplex/AP viscoelastic gel composed of polyethylene oxide (PEO), which is very similar to PEG but has a different molecular weight, and carboxymethylcellulose. Steroids have been used to prevent adhesions and can be administered in a number of ways, including systemically before, during and after surgery; intraperitoneally during surgery; and via hydrotubation postoperatively. Other pharmacological agents used to prevent adhesions include noxytioline, an antibacterial agent; promethazine, an antihistamine; and reteplase, a thrombolytic drug, all of which are instilled intraperitoneally; as well as heparin, an anticoagulant used for intraoperative irrigation. A nasal gonadotrophin-releasing hormone agonist (GnRHa) has also been used preoperatively and postoperatively. How the intervention might work Hyaluronic acid is a major component of many body tissues and fluids, where it provides physically supportive and mechanically protective roles (Johns 2001). PEG is a polymer; when the two PEG-containing liquids are sprayed simultaneously, they form a cross-linked gel. Gels are thought to decrease adhesion formation mainly by preventing denuded tissues from touching. Steroids and antihistamines (e.g. promethazine) act as immunomodulating agents and were used in the belief that they promote fibrinolysis during healing, without hindering the healing process. GnRHa may work by decreasing oestrogen-related growth factors and promoting fibroblasts. Fluid agents such as icodextrin and dextran work through the act of hydroflotation, whereby the fluid separates raw opposing surfaces until the healing process has been completed. Fluid agents are believed to remain in the peritoneal cavity for several days, which is considered a sufficient length of time, given that adhesions form within eight days of surgery (Diamond 2001; Hosie 2001). Why it is important to do this review Adhesiolysis is the only available treatment for adhesions, although controversy regarding its efficacy is ongoing (Hammoud 2004). The focus of adhesion management is now prevention. Intraperitoneal adhesions are associated with considerable co-morbidity and have large economic and public health repercussions. They are the most common complication of gynaecological surgery, forming in 50% to 100% of women (dizerega 1994). Women present with the secondary effects of adhesions including dyspareunia, subfertility, bowel obstruction and chronic pelvic pain, although the latter has a controversial association with adhesions, as no correlation with extent of adhesions and severity of pain is apparent. 6

10 Nevertheless, these consequences can greatly decrease a woman s well-being and require further surgery. Subsequent surgery in women with adhesions is more difficult, often takes longer and is associated with a higher complication rate. It is estimated that in the first year after lower abdominal surgery, the cost of adhesionrelated readmissions in the UK is 24.2 million, which increases to 95.2 million over the subsequent nine years (Wilson 2002). The Surgical and Clinical Adhesions Research (SCAR) study found that 5% (n = 245) of readmissions 10 years after open gynaecological surgery were due to adhesions (Lower 2000; Lower 2004). An English study estimated that the National Health Service (NHS) could save 700,000 per year if an antiadhesion agent that reduced adhesions by 25% and cost 110 was used or, at worst, that this approach would be cost-neutral (Cheong 2011). O B J E C T I V E S To evaluate the role of fluid and pharmacological agents used as adjuvants in preventing formation of adhesions after gynaecological surgery. M E T H O D S 1. Hydroflotation agent versus no hydroflotation agent. 2. Gel agent versus no treatment. 3. Gel agent versus hydroflotation agent when used as an instillant. 4. Steroid (including systemic, intraperitoneal, preoperative and postoperative) versus no steroid (or placebo). 5. Intraperitoneal noxytioline versus no noxytioline (or placebo). 6. Intraperitoneal heparin versus no heparin (or placebo). 7. Systemic promethazine versus no promethazine (or placebo). 8. GnRHa versus no GnRHa (or placebo). 9. Reteplase plasminogen activator versus no reteplase plasminogen activator (or placebo). 10. N,O-carboxymethyl chitosan versus no N,O-carboxymethyl chitosan (or placebo). Types of outcome measures We decided to alter outcomes slightly from the previous version of the review, so that the primary outcomes focus on what is most important to the participants rather than on adhesion formation, which has little correlation with symptoms experienced. A variety of adhesion assessment measures were included as secondary outcomes to enable maximum study inclusion. Criteria for considering studies for this review Types of studies Published and unpublished randomised controlled trials (RCTs) investigating the use of fluid and pharmacological agents to prevent adhesion formation after gynaecological surgery were eligible for inclusion. Non-randomised trials and those considered to be at high risk of bias for sequence generation or allocation concealment were excluded. Studies using a cross-over design were excluded. Types of participants Female participants in any age group who underwent pelvic surgery (by laparoscopy or laparotomy). Studies investigating adhesion prevention in non-gynaecological specialities were not included. Types of interventions Interventions were grouped together for meta-analysis according to physical state and main mechanism of action: hydroflotation agents (including dextran, 4% icodextrin solution), gel agents (including SepraSpray, SepraCoat, Hyalobarrier gel, Intergel, CoSeal, SprayGel and Intercoat) and pharmacological agents. The following comparisons were made. Primary outcomes 1. Pelvic pain (improvement/worsening/no change in pain at second-look laparoscopy (SLL)), independent of the method used to assess pelvic pain. 2. Live birth rate, as defined by the individual study. Secondary outcomes 3. Improvement in adhesion score at SLL, recorded on whichever scale the study authors used, but with preference given to the modified American Fertility Society (mafs) score. 4. Worsening in adhesion score at SLL, recorded on whichever scale the study authors used, but with preference given to the mafs score. 5. Adhesions at SLL. 6. Mean adhesion score at SLL per participant, recorded on whichever scale the study authors used, but with preference given to the mafs score. 7. Clinical pregnancy rate as defined by the individual study. 8. Miscarriage rate, defined as loss of pregnancy before 24 weeks gestation. 9. Ectopic pregnancy rate. 10. Improvement in quality of life (QoL) at SLL, recorded on whichever scale the study authors used, but with preference given to Short Form (SF)-36. 7

11 11. Adverse outcomes, local and systemic, thought to be due to the antiadhesion agent, in studies stating this as one of their outcomes, as opposed to observation. Articles were included independent of the adhesion scoring method used. Articles that met the inclusion criteria but did not report any of the outcomes considered within this review were included within the qualitative analysis. Search methods for identification of studies This is an update of the review by Metwally et al. published in The Menstrual Disorders and Subfertility Group (MDSG) Specialised Register of Controlled Trials, the Cochrane Central Register of Controlled Trials (CENTRAL) and citation indexes were searched using a search strategy designed by the MDSG Trials Search Co-ordinator. No restriction on language was applied. See the Review Group module for additional details on the make-up of the Specialised Register. Electronic searches Electronic databases were searched using Ovid software: MED- LINE (1950 to April 2014), MDSG database (inception to April 2014), EMBASE (1980 to April 2014), CENTRAL (inception to April 2014), PsycINFO (1806 to April 2014) and the Cumulative Index to Nursing and Allied Health Literature (CINAHL) (1982 to April 2014). The clinical trials databases International Clinical Trials Registry Platform (ICTRP) and clinicaltrials.gov were searched from inception to February See Appendix 1; Appendix 2; Appendix 3; Appendix 4; Appendix 5; Appendix 6; Appendix 7; and Appendix 8, Searching other resources Grey literature was handsearched, specifically, abstracts presented at meetings of the British Society of Gynaecological Endoscopy, the European Society of Gynaecological Endoscopy, the American Association of Gynecological Laparoscopists and the British Fertility Society. Reference lists of included studies were also searched. Data collection and analysis Selection of studies Three review authors (GA, FM, DI) independently performed an initial screen of titles and abstracts to assess trials for suitability of inclusion in accordance with the eligibility criteria. FM and DI independently examined the full-text articles and abstracts to confirm eligibility. If necessary, investigators were contacted to obtain further information. Discrepancies were settled by consensus by GA and AW. Data extraction and management Two review authors (FM, DI) independently extracted the data. Data were transcribed onto a Microsoft Word data collection form designed for this review before they were entered into RevMan. The statistical package Metaview of RevMan 5.1, provided by The Cochrane Collaboration, was used to analyse and synthesise data. Study authors were contacted for further information as required. If no reply was received and the information was related to bias, this was denoted as unclear; if the information required was statistical and prevented inclusion in the meta-analysis, the study was not included in that outcome analysis, although it was still considered an included study. Disagreements were resolved by consensus by GA and AW. Assessment of risk of bias in included studies The risk of bias of all studies deemed eligible was assessed independently by two review authors (FM, DI). These included allocation (random sequence generation and allocation concealment); blinding of participants, personnel and outcome assessors; incomplete outcome data; selective reporting and other biases. Disagreements regarding interpretation of data were settled by consensus by GA and AW. The quality of trials was assessed as recommended by the risk of bias tool in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) and was entered into the risk of bias table. Measures of treatment effect The odds ratio (OR) was used for dichotomous data (e.g. number of women with worsening adhesion score). The standardised mean difference (SMD) was used for continuous measures that used different scales (e.g. mean adhesion score at SLL). When the same scale was used, the mean difference (MD) on this specific scale was used. We presented 95% confidence intervals (CIs) for all outcomes. Unit of analysis issues The included primary studies were analysed per woman. Studies that used an internal control were excluded and have been listed as such. Dealing with missing data Investigators were contacted to request missing data. If data were insufficient for inclusion of the study in a particular analysis, it was not included. Assessment of heterogeneity The Chi 2 test was performed and the I 2 statistic calculated to determine significant heterogeneity. An I 2 measurement > 30% 8

12 was considered moderate heterogeneity, > 50% substantial heterogeneity and > 70% high heterogeneity. When significant heterogeneity was identified, the cause was explored, and a sensitivity analysis was performed using the randomeffects model. This was highlighted in the results section, and any variation in the direction of effect was noted. A subanalysis comparing the effects of antiadhesion agents on de novo adhesions versus re-formed adhesions would have been performed if sufficient data were available. Assessment of reporting biases In consideration of the difficulty of detecting and correcting publication bias and other reporting bias, we aimed to minimise the impact by ensuring that a robust and comprehensive search was performed. We planned to create a funnel plot to assess the risk of reporting bias if 10 or more studies were included in a metaanalysis. Data synthesis Statistical analysis was performed in accordance with guidelines developed by The Cochrane Collaboration. Data from the primary studies were combined in RevMan using the fixed-effect model. An increase in OR or SMD or MD was indicated to the right of the central line of the forest plot; a decrease was indicated to the left of the central line. Whether this favoured treatment or no treatment depended on the outcome analysed, but the axes were labelled accordingly. Subgroup analysis and investigation of heterogeneity Sensitivity analysis Sensitivity analysis was performed to determine whether the results were robust to decisions made regarding eligibility of the studies and analysis. If a study was considered to have a high risk of bias, or an apparent outlier was identified, the reason for the significant heterogeneity was investigated, as to whether this was believed to be clinical or methodological, and analysis was conducted to evaluate whether inclusion of the study significantly affected the results. Results of the sensitivity analysis are reported in the Risk of bias in included studies subsection of the results section. R E S U L T S Description of studies Results of the search Forty-four studies were identified as potentially eligible for inclusion. Twenty-nine studies were included. For a summary of each included study, see the section Characteristics of included studies. Reasons for study exclusion are detailed in the Characteristics of excluded studies section. For details of the screening and selection process, see Figure 1. 9

13 Figure 1. Study flow diagram. Included studies Study design and setting Of the 29 included studies, 19 were conducted at multiple centres and 10 at a single centre. Nine were conducted in the USA, six in Europe, two in the USA/Europe, two in the Netherlands, three in Australia, one in Sweden, two in Italy and one in Germany/ Canada/Netherlands/Antilles; three studies did not state their location. Results of nine of the included trials could not be entered into the meta-analysis because the data were not reported in an appropriate format. In some cases, the study authors used different ways of assessing adhesions, such as reporting only individual sections of the mafs, as in Hellebrekers 2009 and Diamond 2003, or the adhesion area (cm 2 ), as in Coddington Another reason why studies could not be entered was that complete statistical data were not published, for example, Thornton 1998 and Rosenberg 1984 did not report standard deviations (SDs) or standard errors of the mean (SEMs), and although Fossum 2011 reported the outcomes we were examining, results were displayed on a graph without actual numbers stated at any point in the text. DiZerega 2007 was not entered into the meta-analysis, as investigators reported only the effect that the antiadhesion agent had on AFS endometriosis score, and as the results were presented per adnexa, not per participant. Thus 20 trials were involved in the meta-analysis. Fifteen studies stated that they received commercial funding. Participants A wide variety was noted in the number of participants in each study, with participant numbers ranging from 10 to 203 in the intervention group and from 10 to 199 in the control group. All 10

14 participants were women undergoing a gynaecological procedure who had had a second-look laparoscopy. Reasons for surgery included pelvic inflammatory disease (PID), endometriosis, adhesions, fibroids, pelvic pain, pelvic mass, endometrioid cysts and infertility assessment and treatment (e.g. tubal surgery). Interventions The numbers of studies entered into the meta-analysis for each comparison are as follows. 1. Seven studies compared hydroflotation agents versus no hydroflotation agents. A distinction was made between hydroflotation agents (e.g. dextran, 4% icodextrin, SepraCoat) designed as antiadhesion agents and liquids such as saline, which was often used as a control and is not considered a hydroflotation agent in this review. 2. Five studies compared gel agents versus no gel agents. 3. Two studies compared hydroflotation agents versus gel agents. 4. Four studies compared steroids versus no steroids. 5. One study compared noxytioline versus no noxytioline. 6. One study compared heparin versus no heparin. 7. One study compared promethazine versus no promethazine. No studies that evaluated GnRHa, reteplase plasminogen activator or N,O-carboxymethyl chitosan could be included in the metaanalysis. Outcomes Two studies did not assess adhesions (Rose 1991; Sites 1997). DiZerega 2007 and Lundorff 2005 presented the results per adnexa. Primary outcomes One of 29 studies examined pelvic pain (Brown 2007). Three of 29 studies examined live birth rate (Jansen 1985; Larsson 1985; Rock 1984). Secondary outcomes Of 29 studies, 11 examined improvement in adhesion score at SLL (Adhesion SG 1983; Brown 2007; dizerega 2002; Jansen 1985; Jansen 1988; Jansen 1990; Johns 2001; Larsson 1985; Lundorff 2001; Mettler 2004; Young 2005). Of 29 studies, nine examined the number of participants with worsening adhesion score at SLL (dizerega 2002; Jansen 1985; Jansen 1988; Jansen 1990; Johns 2001; Lundorff 2001; Mettler 2004; Querleu 1989; Young 2005). Of 29 studies, 10 examined adhesions at SLL (Adhesion SG 1983; Diamond 1998; dizerega 2002; Jansen 1985; Johns 2001; Lundorff 2001; Mais 2006; Mettler 2004; Pellicano 2003; Ten Broek 2012). Of 29 studies, seven examined the mean adhesion score at SLL per participant (Adhesion SG 1983; Brown 2007; Larsson 1985; Lundorff 2001; Mais 2006; Ten Broek 2012; Trew 2011). Of 29 studies, five examined the clinical pregnancy rate. All participants in these studies were actively seeking pregnancy during the study time period (Adhesion SG 1983; Jansen 1985; Larsson 1985; Querleu 1989; Rock 1984). None of the 29 studies examined the miscarriage rate. Of 29 studies, four examined the ectopic pregnancy rate (Jansen 1985; Larsson 1985; Querleu 1989; Rock 1984). None of the 29 studies examined QoL. Of 29 studies, 28 examined adverse outcomes. Rosenberg 1984 was the only study that did not examine adverse outcomes. Excluded studies Nine studies were excluded. Johns 2003 used an internal control. Diamond 2011 and Tulandi 1991 used internal controls, which was not explicitly stated in the abstract. Two studies were interim reports (Mettler 2003(a); Mettler 2003(b)) and the final report was included. One trial was not randomised (Tsuji 2005), and one study did not state that it was randomised (Pellicano 2005), although it appeared to include the same study group as was used in Pellicano This was not explicitly stated in the methods, nor was the fact that the study was randomised. Thus Pellicano 2005 was excluded. One study was excluded because it was quasirandomised (Swolin 1967). Tulandi 1985 reported the effect of the agent on blood indices, not on adhesions. This study was included in the original review but has been excluded because investigators used an external control. Studies awaiting classification Three studies sit in the awaiting classification section (Hudecek 2012; Litta 2013; Tchartchian 2009) pending publication of sufficient data to allow their inclusion. Risk of bias in included studies The risk of bias for each included study can be seen in the Characteristics of included studies section. Figure 2 presents a summary of risk of bias of all included studies. Figure 3 depicts the proportions of studies within each judgement for each risk of bias element. 11

15 Figure 2. Risk of bias summary: review authors judgements about each risk of bias item for each included study. 12

16 Figure 3. Risk of bias graph: review authors judgements about each risk of bias item presented as percentages across all included studies. Allocation Sequence generation No studies were at high risk of sequence generation bias. Seventeen studies adequately explained an appropriate method of sequence generation and were thus deemed at low risk. Twelve studies described the methods of random sequence generation inadequately and were at unclear risk. Allocation concealment No studies were at high risk of allocation concealment bias. Eleven studies were at low risk of allocation concealment bias, as the authors described an acceptable method of allocation concealment. Eighteen studies did not provide sufficient information on allocation to permit a judgement. Blinding Six studies did not provide sufficient information on blinding to permit a judgement. Three studies blinded only the participant (Coddington 2009; Mettler 2008; Rose 1991). Five studies were double-blinded (i.e. both participant and operating surgeon were blinded) (Brown 2007; Mettler 2004; Pellicano 2003; Querleu 1989; Rosenberg 1984). Ten Broek 2012 stated that the study was single-blinded (participant), although the surgeon performing the initial surgery was unaware of allocation until the end of the initial procedure after the adhesions were scored, and the second-look laparoscopy surgeon was blinded. The remaining 14 studies blinded the participant and the operating surgeon and used an independent blinded reviewer to assess videos or diagrams obtained through the second-look laparoscopy (Diamond 1998; Diamond 2003; dizerega 2002; DiZerega 2007; Fossum 2011; Hellebrekers 2009; Jansen 1985; Jansen 1988; Johns 2001; Larsson 1985; Lundorff 2001; Mais 2006; Trew 2011; Young 2005). Incomplete outcome data Two studies (Rosenberg 1984; Thornton 1998) were considered at high risk of attrition bias, as neither study reported SDs or SEMs. Twenty-two studies were at low risk for attrition bias. Five studies did not provide sufficient information to reveal attrition bias; consequently the risk of attrition bias was unclear. Selective reporting One study (Mettler 2008) was at high risk for reporting bias. The authors of the study decided in hindsight to change the primary outcome scoring method from the total mafs score, as stated in the original protocol, to the mafs of the posterior uterus, as discussed during data analysis. Consequently, a sensitivity analysis was conducted and found that excluding Mettler 2008 made no difference to the direction of treatment effect. Thus the study was excluded from analysis. Twenty-five studies were at low risk for reporting bias, and three studies did not provide sufficient information to allow judgement on reporting bias risk. 13

17 Other potential sources of bias Two studies were identified as having other sources of bias that were unclear (Jansen 1988; Ten Broek 2012). A potential source of bias in Jansen 1988 was that the practice of adding hydrocortisone sodium succinate to the irrigation solution was stopped after 46 participants had received it because a possible detrimental effect was reported in an earlier study. These 46 participants were still included in the analysis. The study by Ten Broek 2012 was prematurely ended due to financial and organizational reasons. During the conduct of the study, the clinical trial insurance unexpectedly required a separate fee for both laparoscopic procedures in each patient ; this study was still included. No potential sources of bias were identified in the other 27 included studies. Effects of interventions See: Summary of findings for the main comparison Hydroflotation agents vs no hydroflotation agents for adhesion prevention after gynaecological surgery; Summary of findings 2 Gel agents vs no treatment for adhesion prevention after gynaecological surgery; Summary of findings 3 Gel agents compared with hydroflotation agents when used as an instillant for adhesion prevention after gynaecological surgery; Summary of findings 4 Steroids (any route) vs no steroids for adhesion prevention after gynaecological surgery; Summary of findings 5 Intraperitoneal noxytioline vs no treatment for adhesion prevention after gynaecological surgery; Summary of findings 6 Intraperitoneal heparin solution vs no intraperitoneal heparin for adhesion prevention after gynaecological surgery; Summary of findings 7 Systemic promethazine vs no promethazine for adhesion prevention after gynaecological surgery 1. Hydroflotation agents versus no treatment Primary outcomes 1.1 Pelvic pain One study (Brown 2007) examined the effect of a hydroflotation agent (4% icodextrin) on pelvic pain and found no evidence of a difference compared with saline (OR 0.65, 95% CI 0.37 to 1.14, P value 0.13, one study, 286 participants, moderate-quality evidence). See Analysis Live birth rate No evidence of a difference between groups was seen (OR 0.67, 95% CI 0.29 to 1.58, P value 0.36, two studies, 208 participants, I 2 = 0%, moderate-quality evidence) (Jansen 1985: dextran vs Hartmann s; Larsson 1985: dextran vs saline). See Analysis 1.2 and Figure 4. Figure 4. Forest plot of comparison: 1 Hydroflotation agent vs no hydroflotation agent, outcome: 1.2 Live birth rate. Secondary outcomes 1.3 Improvement in adhesion score at SLL No evidence of a difference between groups was seen (OR 1.27, 95% CI 0.79 to 2.05, P value 0.32, four studies, 665 participants, I 2 = 38%, moderate heterogeneity, moderate-quality evidence) (Adhesion SG 1983: dextran vs saline; Brown 2007: 4% icodextrin vs saline; dizerega 2002: 4% icodextrin vs saline; Jansen 1985: dextran vs Hartmann s). Heterogeneity was reduced to I 2 = 0% when Jansen 1985 was removed, which consequently meant that a significant difference between groups was seen (OR 1.47, 95% CI 1.03 to 2.10, P value 0.03, three studies, 546 participants); 14

18 however a cause for the heterogeneity was not elucidated, and thus the study remained in the meta-analysis. The only difference that we could discern was the use of Hartmann s as a control as opposed to saline by Jansen 1985; however, the review authors believed this difference to be unlikely to cause significant heterogeneity, as the solutions are so similar in composition. See Analysis Worsening in adhesion score at SLL No evidence of a difference between groups was seen (OR 0.28, 95% CI 0.07 to 1.21, P value 0.09, one study, 53 participants, moderate-quality evidence) (dizerega 2002: 4% icodextrin vs saline). With the addition of Jansen 1985, heterogeneity was high in the analysis of worsening adhesion score (I 2 = 79%). As this outcome was poorly defined in Jansen 1985, this study was removed from the analysis; this did not affect the results, as they remained not statistically significant. See Analysis Adhesions at SLL Meta-analysis demonstrated a significant difference in adhesions at SLL, with participants less likely to have adhesions at SLL if they received a hydroflotation agent (OR 0.34, 95% CI 0.22 to 0.55, P value < , four studies, 566 participants, I 2 = 0%, high-quality evidence) (Adhesion SG 1983: dextran vs saline; Diamond 1998: SepraCoat vs phosphate-buffered saline (PBS); dizerega 2002: 4% icodextrin vs saline; Jansen 1985: dextran vs Hartmann s). See Analysis 1.5 and Figure 5. Figure 5. Forest plot of comparison: 1 Hydroflotation agent vs no hydroflotation agent, outcome: 1.5 Number of participants with adhesions at second-look laparoscopy. 1.6 Mean adhesion score at SLL per participant No evidence of a difference between groups was seen (OR -0.06, 95% CI to 0.09, P value 0.44, four studies, 722 participants, I 2 = 0%, high-quality evidence) (Adhesion SG 1983: dextran vs saline; Brown 2007: 4% icodextrin vs saline; Larsson 1985: dextran vs saline; Trew 2011: 4% icodextrin vs saline). See Analysis Clinical pregnancy rate No evidence of a difference between groups was seen (OR 0.64, 95% CI 0.36 to 1.14, P value 0.13, three studies, 310 participants, I 2 = 0%, moderate-quality evidence) (Adhesion SG 1983: dextran vs saline; Jansen 1985: dextran vs Hartmann s; Larsson 1985: dextran vs saline). See Analysis Miscarriage rate This was not assessed by any study. 1.9 Ectopic pregnancy rate No evidence of a difference between groups was seen (OR 0.35, 95% CI 0.06 to 1.85, P value 0.21, two studies, 50 participants, I 2 = 5%) (Jansen 1985: dextran vs Hartmann s; Larsson 1985: dextran vs saline). See Analysis Quality of life This was not assessed by any study Adverse outcomes No adverse outcomes were reported. 2. Gel agents versus no treatment Primary outcomes 15