Interventions for late trabeculectomy bleb leak
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
(1) To compare any interventions for late onset bleb leaks after trabeculectomy with conservative management consisting of observation with or without prophylactic topical antibiotic therapy.
(2) To compare surgical treatment covering the leak with tissue with alternative interventions mentioned above.
Background
Description of the condition
Trabeculectomy is a surgical procedure for glaucoma, in which a guarded fistula is created. This fistula allows aqueous humour to escape from the anterior chamber to the subconjunctival space, where the fluid is reabsorbed through venous capillaries and lymphatics. The intervention was first described in 1968 by Cairns (Cairns 1968). Today, it is the most commonly performed surgical intervention for patients with glaucoma which has not been sufficiently controlled with medication or laser treatment or both.
The accumulation of aqueous humour in the subconjunctival space forms what is called a filtering bleb. The filtering bleb may have different morphologic characteristics such as diffuse or localised, with a thin or thick wall, with a single or multiple cavities, etc. At this level the conjunctiva is the only barrier preventing the entry of infectious microorganisms into the eye.
An increased wound healing response with scarring of the episcleral tissue is the main reason for failure of filtration surgery. To enhance the success rate of filtration surgery wound healing modifying agents such as 5‐fluorouracil (5‐FU) and mitomycin C (MMC) are widely used (Beckers 2003; FFSSG 1996; Hagiwara 2000; Kim 1998; Mietz 1998; Wilkins 2005; Wormald 2001; Wudunn 2002). These antifibrotic agents alter the morphologic and histologic appearance of filtering blebs (Crowston 2004; Khaw 1992). While filtering blebs after trabeculectomy without antimetabolites showed a normal epithelium and loosely arranged subepithelial connective tissue (Addicks 1983), the examination of filtering blebs after the use of 5‐FU or MMC showed a decreased epithelial thickness, decreased number of goblet cells as well as decreased vascularity of the subepithelial tissue (Anand 2006; Francis 2005; Mietz 1996; Shields 1993; Sihota 2000). Additionally a dysfunctional conjunctival barrier with transconjunctival seepage without point leak can occur (Sihota 2000).
These changes may be associated with late onset filtering bleb leaks that can cause severe postoperative complications such as hypotony and bleb related endophthalmitis (DeBry 2002; Greenfield 1998; Matsuo 2002).
Between 1998 and 2004 33, 697 trabeculectomies were performed alone in the UK (Fraser 2006). Especially after the use of antifibrotic agents, the risk of developing point leak in avascular filtering blebs is high with a reported probability of 1.8, 14.6 and 13.6% respectively at 24 months after surgery (Anand 2006; Belyea 1997; Bindlish 2002).
Late bleb leaks typically appear in localised, thin‐walled, avascular filtering blebs. Late leakage of aqueous humour can either be focal or diffuse (e.g. transconjunctival leakage). It is detected by the Seidel test, painting the conjunctiva with a moistened fluorescein strip, cobalt‐blue light filtered illumination and slit‐lamp bio‐microscopy.
Although a late bleb leak may be asymptomatic, it is probably a major risk factor for bleb related infections (Belyea 1997; Greenfield 1996; Mac 2003; Phillips 1994; Soltau 2000). Direct access of tear fluid into the anterior chamber via a leaking filtering bleb has been documented (Gollamudi 1993). Bleb‐related ocular infections can affect initially only the filtering bleb or may involve the whole eye i.e. endophthalmitis, which is a severe and potentially devastating complication. Another potentially sight threatening consequence of late onset filtering bleb leak is hypotony (Newhouse 1982). If a late bleb leak is detected, interventions are typically indicated to halt the leakage and to prevent further complications, even when the patient is asymptomatic at the moment of diagnosis.
Description of the intervention
Different procedures are described to treat late onset bleb leaks. Conservative management with topical antibiotics can be attempted as some bleb leaks may heal spontaneously. Surgical procedures reported in the literature are very diverse and include conjunctival advancement or transplantation (Budenz 1999; Burnstein 2002; Harris 2000; Schnyder 2002; Wilson 1994), scleral or corneal patch (Halkiadakis 2005; Kosmi 1997; Mistelberger 2001), amniotic membrane transplantation (Budenz 2000; Kee 2002), autologous blood, serum or fibrin glue injection (Asrani 1996; Burnstein 2001; Kajiwara 1990; Leen 1995), and nonsurgical interventions including topical application of autologous serum drops (Matsuo 2005), application of cyanoacrylate tissue adhesive (Weber 1989; Zalta 1991) as well as the use of laser treatment (Baum 1993; Hennis 1992).
How the intervention might work
All interventions have the main goal of closing the leak and, in addition, they should preserve the function of the trabeculectomy. Interventions to treat late onset bleb leaks may be divided in two different categories according to their nature and mechanism of action.
(1) Surgical procedures: the defect in the conjunctiva is covered surgically by transplanting or advancing tissue such as conjunctiva, sclera, cornea or amniotic membrane. These interventions might be used alone or in combination with each other. This type of intervention has the aim to seal the defect immediately.
(2) "Alternative treatments": any other procedures will be included such as injection of fibrin glue or autologous blood, application of cryotherapy or laser treatment to the leaking bleb. These treatments could mechanically close the fistula or induce a wound healing response in the subconjunctival space or both. The topical application of substances that promote wound healing such as autologous serum eye drops is also allocated to this category.
Why it is important to do this review
Late onset bleb leak is a complication of trabeculectomy that may lead to severe ocular infection and hypotony. The frequency of such complications may be increasing with the widespread use of antifibrotic agents. Various treatments have been reported in the literature (seeTable 1) and a systematic review is needed to identify and if appropriate summarise any evidence of effectiveness of the different interventions; if possible identifying which is the most effective.
| Intervention | Studies |
| conjunctival advancement | Budenz 1999; Burnstein 2002; Harris 2000; Mandal 2001; Wadhwani 2000 |
| free conjunctival patch | |
| topical autologous serum | |
| autologous blood injection | |
| fibrin glue injection | |
| full thickness scleral graft | |
| corneal graft | |
| amniotic membrane graft | |
| cyanoacrylate tissue adhesive | |
| argon laser treatment |
Objectives
(1) To compare any interventions for late onset bleb leaks after trabeculectomy with conservative management consisting of observation with or without prophylactic topical antibiotic therapy.
(2) To compare surgical treatment covering the leak with tissue with alternative interventions mentioned above.
Methods
Criteria for considering studies for this review
Types of studies
We will include randomised controlled trials (RCTs) and quasi‐randomised controlled trials. In the absence of any RCTs we will discuss non‐randomised comparative studies but not case series and no meta‐analysis will be undertaken.
Types of participants
We will include people with a late‐onset (i.e. more than three months after glaucoma surgery) bleb leak after trabeculectomy, associated or not with lens extraction or application of antifibrotic agents or both. The bleb leak may be symptomatic or asymptomatic, resulting from a localised defect (pinpoint leak) or a diffuse, trans‐conjunctival seepage.
Patients with late onset bleb leak after other types of glaucoma surgery such as deep sclerectomy, viscocanalostomy or implantation of drainage devices will be excluded.
Participants can be asymptomatic, in whom the condition of a leaking filtering bleb is detected by chance, as well as symptomatic with tearing or reduced vision due to hypotony or both.
Patients with a history of blebitis or endophthalmitis in whom the infection has been controlled will also be included. However, leaks due to a trauma (responsible for the conjunctival defect) will be excluded.
There will be no restriction with respect to age, gender, race, co‐morbidities or number of participants.
Types of interventions
Any intervention designed to close a bleb leak. Conservative management (control group) will consist of observation with or without topical lubrication or topical antibiotics or both.
Possible interventions are:
(1) conjunctival advancement or transplantation;
(2) scleral or corneal patch;
(3) amniotic membrane transplantation;
(4) autologous blood, serum or fibrin glue injection;
(5) topical applications of autologous serum drops;
(6) application of cyanoacrylate tissue adhesive;
(7) laser treatment.
Types of outcome measures
Primary outcomes
(1) Sealing of the bleb leak confirmed by a negative Seidel's test after one month, one year and five years.
Secondary outcomes
(1) Recurrence of bleb leak within one year and five years.
(2) Need for further intervention to control bleb leak within one year and five years.
(3) Any surgical interventions to control glaucoma within one year and five years post treatment.
(4) Intraocular pressure and number of antiglaucoma medications at one year and five years after treatment
(5) Visual acuity as measured by logMAR method before bleb leak and before treatment as well as at one and five years follow up.
Although we have set ideal follow‐up times we will include trials with any length of follow up.
Adverse outcomes
The following adverse effects will be of interest.
(1) Infection.
(2) Loss of pressure control (defined by need of additional surgical treatment).
(3) Loss of visual acuity defined as doubling of logMAR compared with both, before leak and before treatment
Additional outcomes
If economic or quality of life data are available from the selected studies these outcomes will be mentioned in the discussion section.
Search methods for identification of studies
Electronic searches
We will search the Cochrane Central Register of Controlled Trials (CENTRAL) (which contains the Cochrane Eyes and Vision Group Trials Register) in The Cochrane Library, MEDLINE and EMBASE. There will be no date or language restrictions in the electronic search for trials.
See: Appendices for details of search strategies for each database.
Searching other resources
We will search the reference lists of the studies included in the review for information about further trials. We will not handsearch conference proceedings or journals specifically for the review.
Data collection and analysis
Selection of studies
Two authors independently will assess the titles and abstracts of all reports identified by the electronic and manual searches. Each report will be labelled A (definitely exclude), B (unsure), or C (definitely include). Full text articles of abstracts labelled as 'unsure' will be reassessed according to the inclusion criteria for this review. Studies labelled 'definitely exclude' will be excluded from the review. We will contact the authors of studies labelled 'unsure' for further clarification. Studies labelled as 'definitely include' will be assessed for methodological quality. We will resolve any differences between the two authors by discussion.
Data extraction and management
Both authors will be extracting data using a paper form developed by Cochrane Eyes and Vision Group. Both authors will enter data into RevMan.
Assessment of risk of bias in included studies
Two authors independently will assess the included studies for sources of systematic bias according to the guidelines in Section 6 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2006). The studies will be evaluated for the following criteria: allocation concealment (selection bias), masking of outcome assessors (detection bias), and rates of follow up and intention‐to‐treat (ITT) analysis (attrition bias).
(a) Allocation concealment will be reported as 'adequate', 'inadequate' or 'unclear'. Any reasonable method of allocation concealment (Higgins 2006) will be considered to be 'adequate'. If the adequacy of allocation concealment is unclear from the trial report we will contact the primary investigators for clarification. If they do not respond within six weeks we will classify the study based on available information and update it as more information becomes available.
(b) Masking of outcome assessors will be noted. Masking of investigators and participants might not be possible with the interventions being examined and will not be assessed.
(c) Rates of follow up, reason for loss to follow up and analysis by the principle ITT will be examined. We will consider a trial to have been analysed by ITT if it analysed participants as randomised and included participants for whom no outcome measurements were made and those who received only part or none of the intended treatment.
If we detect that a study has performed available case analysis (without fulfilling the criteria of ITT) we will still perform meta‐analysis on this study.
We will resolve disagreements through discussion and reach a consensus. We will contact the authors of the studies for additional information on issues that are categorised as 'unclear' from information available in the report. If there is a failure to communicate with the primary investigators, or if they do not respond within two months time we will assess the methodological quality based on the available information.
Measures of treatment effect
For the treatment effect dichotomous data are expected. After the intervention the leak is sealed or the leak persists representing a failure of the procedure. Dichotomous outcomes will be summarised as risk ratios (RR) and continuous outcomes will be summarised as a mean difference (MD). Standardised mean difference will be calculated when outcomes are measured on different scales.
Unit of analysis issues
To address the unit of analysis issue identified trials will be carefully assessed to see whether multiple treatment attempts may cause an unit of analysis error or, although a bleb leak in both eyes is a rare condition and it is not expected to identify studies that include such patients, both eyes of a randomised patient received the same or a different treatment.
In case of multiple treatment attempts or both eyes receiving the same treatment we would define the sample size as number of participants (analogue as described for the analysis of cluster randomised trials).
Dealing with missing data
If it is not possible to extract all of the relevant information from the published reports we will contact the investigators to request individual patient data.
Assessment of heterogeneity
Symmetry of the funnel plot will be examined as well as the inconsistency of effect using the I‐squared statistic. If the I‐squared statistic is greater than 50% we will consider it to be a significant role of heterogeneity in explaining the variation between effect estimates of included studies and we will not combine the study for a meta‐analysis. An I‐squared statistic of less than 50% will be used to signify no significant clinical heterogeneity and no funnel plot asymmetry. The outcome of the included trials will then be combined for a meta‐analysis using the random‐effects model. If there is no statistical or clinical heterogeneity, and if the number of trials is fewer than three, a fixed‐effect model will be used so as to avoid reporting less robust effect estimates that may result from random‐effects models in situations with very few trials.
Assessment of reporting biases
To assess reporting biases we will prepare a funnel plot using RevMan to look for signs of asymmetry.
Data synthesis
Data analysis will be done according to the guidelines set out in Section 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2006).
Subgroup analysis and investigation of heterogeneity
In our review the treatment effect may vary with different treatment characteristics and a large number of population characteristics, as different types of interventions are investigated and compared with the number of patients with bleb leak the number of relevant underlying conditions is high. Therefore, it maybe necessary to undertake subgroup analysis. However the number of eligible studies is expected to be very small.
Sensitivity analysis
We will examine the impact of excluding studies with lower methodological quality, unpublished data, and industry funded data in sensitivity analyses.
| Intervention | Studies |
| conjunctival advancement | Budenz 1999; Burnstein 2002; Harris 2000; Mandal 2001; Wadhwani 2000 |
| free conjunctival patch | |
| topical autologous serum | |
| autologous blood injection | |
| fibrin glue injection | |
| full thickness scleral graft | |
| corneal graft | |
| amniotic membrane graft | |
| cyanoacrylate tissue adhesive | |
| argon laser treatment |
