Surgical management of glaucoma: a review


  • William H Morgan FRANZCO PhD,

    Corresponding author
    1. Lions Eye Institute, Centre for Ophthalmology and Visual Science, University of Western Australia, Nedlands, Western Australia, Australia
      Dr William H Morgan, Lions Eye Institute, Centre for Ophthalmology and Visual Science, University of Western Australia, 2 Verdun St, Nedlands, WA 6009 Australia. Email:
    Search for more papers by this author
  • Dao-Yi Yu MD PhD

    1. Lions Eye Institute, Centre for Ophthalmology and Visual Science, University of Western Australia, Nedlands, Western Australia, Australia
    Search for more papers by this author

  • Competing/conflicts of interest: No stated conflict of interest.

  • Funding sources: No stated funding sources.

Dr William H Morgan, Lions Eye Institute, Centre for Ophthalmology and Visual Science, University of Western Australia, 2 Verdun St, Nedlands, WA 6009 Australia. Email:


This review will briefly describe the history of surgical glaucoma treatment and concentrate on the current surgical strategies for managing glaucoma. A discussion of treatments for angle closure, open-angle glaucoma and paediatric glaucoma with an emphasis on drainage surgery are included. The role of cataract surgery is also briefly described. Drainage surgery evolved from peripheral iridectomy and sclerotomy with an increasing understanding of aqueous flow within the eye and the production of a functioning bleb. The current mainstays include trabeculectomy, glaucoma drainage devices as well as goniotomy and trabeculotomy, which have all been in existence for more than 40 years. Their various advantages as well as methods used to minimize their disadvantages, including the antimitotics and case selection are discussed. We finish by discussing the preliminary results of some newer forms of drainage surgery illustrating the energetic search for methods to minimize the problems of hypotony and bleb failure.


Glaucoma is a disease of the optic nerve, which leads to typical optic disc excavation with loss of ganglion cell axons and visual field loss. Its major risk factor is raised intraocular pressure (IOP), which is the only factor currently treatable. Hence, all treatment is directed towards lowering IOP towards a safer level in individual patients. Surgical treatment is required when medical or laser procedures have been used, considered or failed. The basic principles of surgical treatment are to either: (i) improve aqueous flow within the eye, enabling aqueous access to the angle drainage structure; (ii) increase aqueous egress from the eye; and (iii) reduce aqueous production by the ciliary body. Procedures to improve the resilience of the optic nerve have been attempted but without clear documentation of success.1

This review will briefly describe the history of surgical glaucoma treatment and concentrate on the current surgical strategies for managing glaucoma including a somewhat detailed discussion of the more commonly performed procedures. It includes a discussion of treatments for angle closure, open-angle glaucoma and paediatric glaucoma and hence is necessarily brief in order to accomplish an overall summary.


The earliest surgical treatment described in modern times is ascribed to William McKenzie in 1830 from Glasgow, referenced by De Wecker in 1879.2 McKenzie performed a sclerotomy, passing a narrow knife through the sclera 1 mm posterior to clear cornea. The aim was to create a junction between the anterior chamber and external space. It is not known what the overall success rate was; however, complications to the ciliary body were described.

Peripheral iridectomy was first described by Albrecht Von Graefe in 1857.3 He stated that it was often curative for glaucoma and that the development of a cystoid scar postoperatively was advantageous. This cystoid scar was probably a bleb, and since those times, the generation and maintenance of a bleb has been a hallmark sign of successful drainage surgery. In the 1870s sclerotomy, was described by De Wecker and Holth, cutting a channel through sclera underneath the conjunctiva to the anterior chamber producing a filtering cicatrix.2

In 1904, Soren Holth from Oslow described iridencleisis, whereby the sclerotomy was performed and iris was incarcerated into this wound beneath the conjunctiva in order to maintain patency through the sclerotomy.4 Full thickness procedures of sclerotomy and iridencleisis remained the procedures of choice until trabeculecomy was described by Cairns in 1968.5 Congenital glaucoma was revolutionized by the description of goniotomy and the development of surgical gonio lens by Barkan in the 1930s.6

In 1969, Anthony Molteno described the first glaucoma drainage tube device, consisting of a plastic tube to plate, which became the forerunner of all such devices.7 Currently, derivatives of trabeculectomy, glaucoma drainage devices (GDDs) and goniotomy have become the mainstays of all glaucoma drainage procedures.

Non-penetrating surgery was first popularized by Krasnov in 1964 with sinusotomy (deep sclerectomy)8 in an effort to reduce postoperative hypotony and other more common complications of full thickness drainage procedures. This has been modified more recently with the addition of a collagen wick-type implant.9

IOP elevation in glaucoma

Most of the aqueous outflow from the eye occurs through the trabecular meshwork, and hence, access to the meshwork is vital for IOP regulation. Trabecular meshwork function may become diseased, tending to block aqueous passage through to the canal of Schlemm and aqueous veins. Normal aqueous humour resistance occurs mainly around the juxta-canalicular meshwork;10 however, it is not known where the bulk of resistance change occurs in primary open-angle glaucoma, although it is assumed to occur in the juxta-canalicular and adjacent trabecular meshwork tissue.

Internal forces within the eye can lead to a force pushing the iris-ciliary body diaphragm or iris-lens diaphragm forward against the trabecular meshwork causing occlusion.11 A combination of these forces due to increasing fluid resistance between the lens and iris, lens and ciliary body, increased lens size and mal position of the ciliary body are the major underlying causes of primary angle closure.12,13 Inflammatory, fibrovascular or metaplastic tissue changes within the angle can lead to contraction of iris across the meshwork. In primary congenital glaucoma, the trabecular meshwork has not cleaved and differentiated normally.

Indications for surgery

Surgical treatment is indicated when the form of glaucoma is clearly defined and medical treatment has failed. This usually includes clear evidence of glaucomatous progression, that is visual field and/or optic disc changes from baseline, or that the IOP is in excess of a logical target considering the patient's life expectancy and other needs. The type of glaucoma surgery must be carefully chosen to fit the particular patient in order to minimize complications and maximize the chances of success.14

Angle-closure glaucoma

Surgical peripheral iridectomy was the mainstay of acute and chronic angle-closure glaucoma treatment until the advent of modern laser peripheral iridotomy.3 It is important to recognize that persistent angle closure is common following laser or surgical iridectomy, and the patient must be kept under regular review, including regular gonioscopy.15 It is important to clarify the cause of the angle closure and initially note whether a pushing force is being applied to the iris and ciliary body, forcing them towards the trabecular meshwork and leading to closure. This is generally due to aberrant anatomical structures and flow relationships within the eye and underlies primary angle closure. Particular attention needs to be made towards diagnosing pupil block in order to justify iridectomy in the first place and secondly noting the presence of plateau iris, lens disproportion or ciliary block.13,16 For example, plateau iris tends to respond better to peripheral iridoplasty than to lens extraction, whereas lens disproportion and ciliary block generally respond well to standard lens extraction. A more complete description of these factors is described by Ng and Morgan.13 Non-primary angle-closure mechanisms include rubeosis iritis17 and uveitis18 both with peripheral anterior synechial closure as well as less common conditions such as irido-corneal endothelial syndrome with metaplastic tissue causing contraction and rarely tumours leading to seeding of the angle or mass effect inducing compression. These types will not be successfully treated by iridectomy or other internal flow enhancing procedures and will usually require drainage surgery.

Cataract surgery

Lens extraction tends to open the angle, eliminating lens disproportion, reducing ciliary block, and to some extent, plateau iris.19,20 In these situations, it is important to assess the angle and determine whether the closure is appositional, or whether there are some peripheral anterior synechiae present. During the cataract surgery, one can expect an increased rate of complications. These range from a shallow anterior chamber with decreased instrumentation and capsular access. The use of high viscosity viscoelastics frequently makes these steps safer and easier. There is less room for lens fragment manipulation, and it is easier for corneal endothelial damage to occur. Additionally, the posterior capsule frequently appears to be under positive pressure, making the likelihood of capsule rupture greater. Rarely, these patients can go into ciliary block on the table, and so the surgeon should be prepared to perform a vitrectomy, zonulectomy and anterior hyaloidectomy if required.21,22 Patients can develop ciliary block, post cataract extraction, and the same corrective surgery may be required. If patients are noted to have peripheral anterior synechiae of recent onset, then goniosynechiolysis should be considered during the lens extraction procedure.20 Various techniques have been described, but we find the easiest technique is to use the irrigating cannula at the end of the procedure whilst removing viscoelastic and gently push the iris away from the corneo–scleral junction in the region of the peripheral anterior synechiae. Results of goniosynechiolysis are mixed but can be favourable, particularly and only when the underlying cause of the angle closure is being treated as well.23,24

Cataract extraction surgery significantly lowers the pressures in all forms of glaucoma, including primary open-angle glaucoma, through mechanisms which are not entirely understood. The average IOP reduction in primary open-angle glaucoma is reported to be from 2–4 mmHg.25 This pressure lowering can be long lasting, although some studies indicate that it may only last a year or so. So, effective cataract surgery can complicate the interpretation of results of drainage surgery where the drainage surgery is combined with cataract surgery. It is likely that some of the effect of combined cataract surgery/drainage surgery is simply due to the removal of the cataract itself.

Drainage procedures

These form the bulk of glaucoma surgery. They are only safe if there is no intraocular force tending to push the lens-iris, lens-ciliary body diaphragm anteriorly. If one performs a drainage procedure in this situation (classically this occurs in ciliary block), then the reduced pressure in front of this force will lead to exacerbation of its effect with the lens-iris diaphragm being pushed further forward, often leading to corneo-lenticular touch and other complications.26

Trabeculectomy is the classic and current mainstay for glaucoma drainage surgery and was originally conceived as a way of shunting aqueous humour through to the canal of Schlemm by excising part of the trabecular meshwork exposing the canal of Schlemm.27 It was quickly observed that a bleb was formed, implying that the fluid drainage was occurring through to the sub-conjunctival space. More recent emphasis has been on performing a simple clear corneal ostomy avoiding dissection of the trabecular meshwork to minimize the risk of ciliary body damage. The aim is to form a safe, functioning bleb.


It is essential that the bleb is safe in the long run and has minimal chance of breaking down or leaking and becoming infected. For this reason, trabeculectomy surgery is generally performed beneath the upper eyelid so the bleb is protected by the lid. Allowing a bleb to form inferiorly increases the risk of bleb exposure with a greater risk of infection.28 It is preferable to attempt the formation of a more posterior bleb to minimize the chance of bleb exposure at the limbus and also to minimize the chance of bleb extension upon the cornea.29


The bleb must be able to collect and disperse aqueous humour. The dispersal of aqueous humour is poorly understood; however, recent evidence suggests that sub-conjunctival lymphatics and formation of aqueous vein connections, as observed in animal models (Fig. 1), are likely to be important in the maturation of human trabeculectomy blebs.30 Bleb morphology is a poor predictor of IOP reduction, although a higher bleb is more likely to result in a lower pressure.31 Most surgeons prefer a modestly elevated, diffuse, relatively large and posterior bleb for reasons of safety. Most observers try to avoid a significantly vascularized bleb. Trabeculectomy bleb failure is commonly ascribed to the process of fibrosis, gradually contracting the bleb down to and flattening it against the scleral flap. This process of fibrosis may occur around the bleb and limit the egress of aqueous humour from the bleb to the surrounding lymphatics and vessels.

Figure 1.

Sequence of video frames (Panels A–I) after injection of fluorescein into the anterior chamber of a rabbit 2.4 years after microfistula drainage surgery. Panel A is a frame recorded immediately after fluorescein injection and shows filling in the anterior chamber (AC). The first appearance of extraocular fluorescein (Panel B, yellow arrow) is a small patch near the distal end of the scleral channel. The extent of fluorescein increases (Panel C), and draining lymphatics become apparent (Panels D–F, wavy arrows). An aqueous vein can be seen located at the distal edge of the bleb (Panels G–I). However, it was difficult to determine whether this vein was a normal aqueous vein superimposed on the bleb, or whether it played a role in draining the bleb. Panel J shows a magnified image of the late phase (Panel I). The conjunctival bleb was small in size and surrounded by a narrow diffusion zone along with a number of drainage vessels. At least two large lymphatic vessels were seen with uneven calibre located each side of the conjunctival bleb and running parallel to limbus (brownish bent arrow). A conjunctival vein was also visible. From figure 15 in Yu et al.30

Surgical technique for trabeculectomy

The technique has remained very similar to the original description by Cairns 43 years ago.5 There have been some stylistic variations with many surgeons initially performing a fornix-based flap and then shifting to a limbal-based flap with the introduction of antimitotics and concerns over wound healing. More recently, there has been a tendency towards shifting back to a fornix-based flap in an effort to encourage more posterior drainage. Other stylistic variations include, whether a triangular or rectangular scleral flap is cut, the length of the flap and the thickness of the flap. A longer flap may tend to encourage more posterior drainage, and a thinner flap may tend to encourage greater flow. None of these factors or style variations have been shown to make any significant difference to longer term outcomes.32 The use of releasable sutures does allow for some titration of pressure in the postoperative period. It is generally safer to aim for a tighter flap during surgery and hence attempt to have a modest pressure reduction down towards 5 mmHg, rather than 0 mmHg, and release a suture if that is not reached. Other attempts at pressure modification in the early postoperative period include laser suture lysis, which unfortunately can be difficult or impossible if there is overlying conjunctival haemorrhage. The other modifiable process is that of the fibrosis itself. Postoperative 5-fluorouracil injections have been shown to improve bleb outcome and reduce IOPs in selected patients.33 These can be given either as a bolus adjacent to the bleb or involving some needling of peri-bleb fibrotic tissue in the later postoperative period.34

Bleb modifying agents include mainly antimitotic drugs such as 5-fluorouracil and mitomycin C. The use of these agents does represent a major advance in achieving more desirable outcomes for patients, particularly those who are at higher risk of trabeculectomy failure. This risk increases with certain racial genetic ancestry such as those with African, East Asian or South Asian decent. Other risks for failure include those with any prior intraocular surgery, uveitis or rubeosis iridis. Eyes which have had significant trauma and particularly eyes that are aphakic are very prone to failure. Younger patients are also more prone to increased fibrosis. So when assessing a patient for trabeculectomy surgery, it is important to assess their risks of bleb failure and your own surgical experience. Younger patients and particularly those from the earlier mentioned racial backgrounds should, in general, receive mitomycin C at the time of surgery. Caucasian or older patients could receive 5-fluorouracil at the time of surgery. Certain patients should not receive any antimitotic agents. These may be elderly Caucasian patients and particularly patients who are on immunosuppressant agents already, for example, those who may have coexistent rheumatoid arthritis treated with systemic steroids and other immunosuppressants. Injections of 5-fluorouracil adjacent to the bleb can be given postoperatively in order to reduce fibrosis and lower IOP. The risks of 5-fluorouracil include corneal epithelial stem cell toxicity with punctuate epithelial erosions leading to epithelial breakdown and stromal thinning.34 Patients being given 5-fluorouracil should have their corneal epithelium carefully monitored, and further injections of 5-fluorouracil should not be given if signs of toxicity are seen. Mitomycin C has very profound effects upon mitosis and can lead to very thin avascular blebs which pose a significant risk of infection.34 It can also inhibit the healing of the scleral flap upon the scleral bed, leading effectively to a full thickness procedure with marked hypotony and attendance complications.

The effectiveness of trabeculectomy

The success rates of trabeculectomy vary between 40–98%.35 Two major reasons exist for this variation. One is that the definition of success is so varied, with there being at least 100 separate definitions of drainage surgery success, so comparing one study with another is often very difficult.35 The other factor is the patient group variation used in different studies due to factors mentioned earlier. Generally accepted effectiveness in appropriately chosen trabeculectomies for low-risk patients is about 90–95% significant pressure reduction for at least several years and 83% 10-year survival.36 The average 5 year survivals for mixed low and higher risk cases are from 6037 to 80% (Fig. 2).38 Please note that definitions of success vary between studies. Early complications of trabeculectomy include hypotony with the development of hypotonous maculopathy, choroidal effusions, shallow anterior chamber and corneal lenticular touch with corneal trauma. The commonest cause of severe hypotonous complications is a leaking trabeculectomy bleb, often through inadvertent cutting of a conjunctival button hole or conjunctival wound dehiscence. It is essential to check the wound and conjunctiva for leaks in the early postoperative period using the Seidel test. Hypotony also occurs in patients who are simply over-draining with a large bleb. In this situation, hypotony will usually resolve spontaneously. However, if there is a conjunctival leak, these complications can persist and cause significant permanent disability. Hence, any conjunctival leak or dehiscence should be closed as soon as possible. More significant is the risk of infection, termed blebitis, in the late postoperative period. This risk approximates 1.2% in Japanese over a mean 3.5 years in patients treated with no antimitotic.39 In a North American population, the risks of endophthalmitis (7.5%) and blebitis (6.3%) over 5 years following mitomycin C trabeculectomy are higher.40 Hence, it is important to carefully select the antimitotic for the particular patient. The most common complication in the late postoperative period is failure with fibrosis of the bleb. Initial signs are usually a rise in pressure, cystic or flattened appearance of the bleb with surrounding scar tissue, which may be hypervascular, tending to wall it off. This can often be treated with 5-fluorouracil needling. If this fails, then medical therapy may need to be reinstituted or a subsequent additional drainage procedure performed.

Figure 2.

Kaplan–Meier probabilities of failure of interventions. Estimated cumulative failure rates (and 95% confidence intervals) at 5 years are 35.5% (30.6–40.4), 42.4% (29.0–55.9) for argon laser trabeculoplasty (ALT) as first and second interventions, respectively; 17.9% (13.9–21.9), 20.7% (13.3–28.2), 34.4% (17.6–51.1) for trabeculectomy as first, second and third interventions, respectively. From figure 1 in The AGIS Investigators.38

Patients need to be carefully informed about the nature of this surgery and in particular the short- and longer term risks. It is essential to inform patients that their vision will not be improved by surgery, and in fact, it could be somewhat worse. Some 7–19% of patients experience a sustained reduction in visual acuity postoperatively.26,41 The rate of cataract formation is increased following drainage surgery and so patients need to be warned of this possibility. They need to be warned that an artificial channel is being created, and that the body will attempt to close it down through the process of fibrosis and that you may need to give injections or other agents to mollify this process.


In 1969, Anthony Molteno first described a silicone tube to polyethylene plate apparatus designed to minimize the problems associated with fibrosis obliteration of blebs.7 The functional principle was that the silicone tube inserted into the anterior chamber and connected to the plate acted as a conduit and diffusional surface area that could not be obliterated. The success of this design has seen the production of similar devices by other groups (Baerveldt, Ahmed, Krupin).


The most common indication for GDD implantation is where trabeculectomy surgery, particularly with an antimitotic agent has failed or where trabeculectomy surgery is most likely to fail. The latter form primary tube surgery indications and include iridocorneal endothelial syndrome and rubeosis iridis, both of which do very poorly with trabeculectomy. Additionally, complicated glaucomas due to chronic uveitis and aphakia generally do better with GDDs than with antimitotic-augmented trabeculectomies. The same essential aims are being achieved by these devices, namely the formation of a safe, functioning bleb. The blebs do tend to be safer following glaucoma drainage device insertion, mainly because they are posterior to the rectus muscle insertions and induce overlying fibrous tissue encapsulation. There is a reduced late risk of infection following drainage tube implantation. The exceptions here are when the wound dehisces or the tube erodes through the conjunctiva.

Glaucoma drainage tube results

Glaucoma drainage tube function tends to last longer than trabeculectomy. A 10-year study of primary tube surgery with some additional risk factors showed qualified success rates of approximately 95%.42 Three-year success rates of 85% when tubes are inserted following cataract surgery or trabeculectomy are reported in a multicentre randomized trial.43 It is however, important to acknowledge that these patients were mainly elderly Caucasians, whereas success rates of 50% at 5 years (Fig. 3) from other high-risk clinical and demographic groups are reported.44

Figure 3.

Kaplan–Meier survival curves for all patients, plotting the cumulative probabilities against time that (i) the intraocular pressure (IOP) remains below 22 mmHg without additional medical treatment (‘complete’ success: CS) (ii) the IOP remains below 22 mmHg with or without additional medical treatment (‘qualified’ success: QS); and (iii) the IOP remains below the target IOP (set as a 30% reduction from the mean preoperative IOP; −30%R) following tube implantation. From figure 2 in Broadway et al.44


The long-term safety of GDDs are generally excellent. The major concern for most surgeons is the risk to the cornea of having a silicone tube passing through or adjacent to the endothelium. In patients with one single tube entry into the anterior chamber, the reported risk of corneal decompensation is 18%.45 Where initial tube surgery has failed due to fibrosis and IOP elevation and a second tube insertion has been performed, this carries a higher risk of corneal decompensation of up to 44%.46 The risk to pre-existing corneal grafts is high, and hence, the concern is in patients with penetrating keratoplasty. In those patients, it is wise to cover them with systemic steroids in the immediate postoperative period. Those patients need to be warned of symptoms of graft rejection also. Most surgeons will try to position the tube as far from the cornea as possible. Pars plana insertion in vitrectomized eyes is thought to reduce corneal decompensation risk but can increase posterior segment complications.47 We do this where possible, particularly in children. We also position the tube between the iris and intraocular lens where possible (Fig. 4).

Figure 4.

Molteno tube inserted between iris and intraocular lens in a subject with iridocorneal endothelial syndrome.

Other complications are similar to those occurring with trabeculectomy. In particular, hypotonous complications can occur in the early postoperative period and again are most likely due to wound dehiscence or leakage through other sites within the conjunctiva. Any conjunctival wound dehiscence or erosion must be treated immediately. A wound breakdown over the plastic tube or plate is a serious complication and usually will not heal without the use of a free or advancement conjunctival graft over a scleral patch graft. The risks of leaving a conjunctival wound dehiscence or wound breakdown are dire. Because the apparatus is plastic, it is not uncommon for epithelium to grow around the plastic in an effort to exteriorize this foreign body. The consequence is that the epithelium will gradually grow down the tube towards and into the anterior chamber and can cause epithelial ingrowth. Additionally, desquamating epithelial cells can pass down the tube and enter the anterior chamber.

Surgical technique variations

With any of these devices, the plate should be sutured behind the level of the rectus insertions. The tube should be covered either by a thick scleral flap or preferably incorporated with a free scleral graft above or below the scleral flap under the conjunctival closure. This extra bulk over the tube minimizes the risk of early or late tube extrusion. The conjunctiva needs to be closed carefully to prevent wound dehiscence. This is extremely important in young children where it is usually impossible to examine them until the next examination under anaesthesia. It is most important to warn patients to not rub their eyes. Rubbing the eye leads to the tube abrading adjacent tissue; that is iris, lens, ciliary body and cornea. It also puts significant stress upon the wound and increases the risks of wound dehiscence. This is very important in young children where the wound simply cannot be observed at routine visits. Antimitotic agents have been used in conjunction with GDDs but have not been shown to make any difference to the long-term pressure control or outcome. Mitomycin C applied to the overlying Tenon's layer has however, been shown to reduce the IOP rise during the hypertensive phase.48 The tube should be occluded for the first 6 weeks postoperatively in order to avoid severe hypotony. This is easily achieved using an absorbable vicryl suture and cutting a small slit in the tube between the vicryl and anterior chamber to allow some temporary drainage.

The pressure will usually fall in the first few weeks then gradually rise until the vicryl suture dissolves at about 6 weeks postoperatively. Then it is likely to fall again and rise as fibrovascular tissue grows around the plate. In younger patients in particular, this often causes the so-called ‘hypertensive phase’ whereby the IOP can rise excessively, necessitating treatment with aqueous suppressants. The hypertensive phase ends as the fibrovascular response matures, usually at 5–12 months post-surgery, with a longer time in younger patients.48

Comparisons between drainage tube devices

There appears to be little significant difference in long-term pressure lowering between the three major devices currently used (Molteno, Baerveldt, Ahmed);49 however, true randomized controlled trials comparing them are few and short, with 2 to 3 years follow-up. The Molteno and Baerveldt are non-valved devices and tend to have a lower IOP in the late postoperative period compared with the Ahmed valve.50 The profile of complications is similar and does not appear to be significantly different between the devices. Bleb encapsulation and a hypertensive phase at 3 months post-sugery appear more likely with Ahmed.51 The 350 mm2 and 500 mm2 Baerveldt tube plate system are associated with a greater risk of strabismus and extraocular muscle movement disorder postoperatively.50 The size of the plate has been thought to be a significant factor in determining postoperative pressure reduction; however, beyond a plate surface area of approximately 200 mm2, there appears to be little difference in pressure reduction.49 The paediatric and early single-plate Molteno tubes appear to produce a higher IOP than the double plate and more recent third-generation larger single-plate tubes. However, there appears to be little difference between the larger Molteno tube plates and Baerveldt tube plates in terms of IOP reduction.50

Modifying bleb function

There is no evidence that postoperative injections of 5-fluorouracil or mitomycin C improve or favourably modify drainage tube device function. Attempts have been made to excise thickened Tenon's layer overlying the plates, but again, this has not been shown to cause significant or long-lasting pressure reduction.46 The use of systemic fibrosis modifying agents (prednisolone, colchicine and non-steroidal anti-inflammatory drugs) may modify bleb maturation in the early postoperative period; unfortunately, there is little evidence, and there are no randomized controlled studies to clearly support their use.42

Paediatric glaucoma drainage surgery

Idiopathic primary congenital glaucoma is the single largest cause of paediatric glaucoma and is identified when there is no other physical abnormality afflicting the patient. When there is another associated physical disorder, this is termed secondary paediatric glaucoma. The commonest cause of the latter is congenital cataract which is strongly associated with subsequent paediatric glaucoma. Many paediatric syndromes are also associated with glaucoma. The primary congenital glaucomas generally respond very well to goniotomy or trabeculotomy52 as well as trabeculectomy.53 Visual stability of 71% at 10 years and 58% at 34 years are reported.53 Short-term IOP success rates of 90% are reported.54 Secondary congenital glaucomas usually respond poorly to these procedures, and so, it is important to make an accurate diagnosis in order to aid prognostication and the optimum choice of surgery.

In paediatric glaucoma where there is significant other ocular abnormality such as Peter's syndrome, persistent hyperplastic primary vitreous and following a congenital cataract removal, a GDD is usually required.55 If the secondary glaucoma occurs before 12 months of age and the associated ocular abnormalities are not too significant, then trabeculectomy or goniotomy can be attempted. In primary congenital glaucoma, successful trabeculectomy or goniotomy is more likely if surgery is done prior to the first 12 months of life.54 After 12 months of age, the success rate dramatically declines. This effect may be due to the reduced plasticity of drainage tissues with ageing.


Goniotomy was the first described and requires a good view of the angle so that the goniotomy knife can be inserted ab interno, across the anterior chamber to engage the trabecular meshwork in order to cut down to the canal of Schlemm.6 Generally, a 120° incision of the meshwork is made. Trabeculotomy requires surgical exposure similar to trabeculectomy wherein a conjunctival and scleral flap are created, then a radial incision is made down to the canal of Schlemm, which needs to be clearly identified. An angled Harms trabeculotome is inserted into the canal of Schlemm. The parallel prong of the probe is kept anterior to the wound, and the device levered centrally, rupturing the inner wall of the canal of Schlemm and the overlying trabecular meshwork, creating a junction between the anterior chamber and canal. This procedure can be performed when the anterior chamber view is poor, which is very useful in some patients with buphthalmos. If the canal of Schlemm cannot be found or seen, then a standard trabeculectomy without antimitotic can be performed.53

Postoperatively, a bleb is frequently observed to last a few months before flattening out and becoming obliterated. However, in the majority of patients, significant pressure reduction remains. It is possible in these patients that junctions between the canal of Schlemm or surgical ostomy and lymphatics or aqueous veins have been created. Unfortunately, pressure reduction in the absence of a bleb occurs only very rarely in adults following surgery.

The outcomes from the three procedures are comparable, and little significant difference exists between trabeculotomy, goniotomy or trabeculectomy in the long term.53,56 However, trabeculotomy success rates of 67%, compared with 54% for trabeculectomy, are reported in a higher risk Arabic population.57 The procedure chosen tends to depend upon surgeon preference as well as patient media clarity.

Secondary congenital glaucoma treatment

Trabeculectomy with mitomycin C has been attempted in many patients with these conditions. However, there is a more recent trend towards using GDDs with better pressure lowering reported.55 The concerns are long term, particularly related to corneal decompensation. If the child has had a vitrectomy, then our practice is to insert the tube through the pars plana to minimize possible corneal contact. With congenital cataract surgery and intraocular lens placement, we will tend to implant the tube either between the iris and intraocular lens or as far posteriorly within the anterior chamber as possible. There is scant long-term data regarding treatments for secondary congenital glaucoma, but the data from the use of GDDs are promising.55

Cyclodestructive procedures

Cyclodestructive procedures using laser or cryotherapy are used to reduce the formation of aqueous humour by the ciliary processes. Most commonly, this is performed with a cyclodiode laser using an 810-nm continuous wave laser. This can be performed as an outpatient or in children under general anaesthetic. More recently, endoscopic cyclodiode laser has been used in some patients. Cyclodiode laser is generally reserved for patients who have failed drainage surgery and has reported pressure lowering success rates of 38 to 55% at close to 2 years.58,59 There have been comparisons of this with the GDDs; however, cyclodiode laser tends to cause more inflammation, and the longer term visual results show that up to 30% can lose two or more lines or vision.60 Poorer visual outcomes from cyclodiode laser compared with drainage devices are demonstrated in several studies.58,59 Generally, cyclodiode laser is used in patients who have failed GDDs or who lack useful vision.


Significant pain can be expected for the first few days after laser and must be explained to the patient. Usually, only 180° of the ciliary body circumference is treated. Usually two or three treatment episodes are required for long-standing pressure control. There is an 8% risk of hypotony at 2 years particularly in neovascular eyes.58,59 If this occurs, then unfortunately, it cannot be successfully treated, and the patient will end up with a phthisical eye. This is the reason for judicious treatment over 180° only at each treatment session. There is a small risk of sympathetic ophthalmitis following cyclodiode laser.61 Curiously, sympathetic ophthalmitis is not reported following uncomplicated trabeculectomy or GDD insertion. Because patients undergoing cyclodiode laser generally have markedly reduced facility of outflow, and cyclodiode laser reduces aqueous production, the diurnal and other variations in aqueous production will lead to increased IOP variation. Hence, following cyclodiode laser, the IOP tends to be brittle, with large fluctuations which can be an additional problem for glaucoma control. For all of these reasons, our own practice is to relegate cyclodiode as the last choice in surgical intervention.

Other drainage procedures

Several procedures are being popularized at present, with suggestions that they may minimize post surgical complications and yet still produce acceptable or excellent long-term pressure reduction. Unfortunately, most of the published follow-up periods are short, being of the order of 1 year or less, and the definitions of success vary greatly, making it difficult to compare these new procedures with each other or with standard drainage procedures. None of them have been shown to be superior to standard trabeculectomy.62

Deep sclerectomy with collagen implant has been reported to have a 56–90% success rate at 1 year, but meta-analysis demonstrates less effective IOP reduction compared with trabeculectomy.63 There are fewer hypotony-related complications reported.63 Bleb formation is noted to occur with deep sclerectomy and can be modified with antimitotics. Viscocanalostomy has a reported success rate of 36–79% at 1 year.64 Metal (EX-PRESS, Alcon, Hunenberg, Switzerland) shunts across the inner corneosclera beneath a trabeculectomy-style scleral flap, replacing hand-cut ostomy, are reported to have up to an 88% success rate at 1 year and 70% success at 5 years (IOP ≤ 18).65 Cyclodialysis cleft formation with the insertion of a metal implant to maintain patency is reported to have some success, but the results are not reported in peer-reviewed literature. Cyclodialysis cleft without implant is reported to have 14% success rate at 1 year.66

Ab-interno procedures, which tend to minimize conjunctival surgical trauma, have been described. Trabecular meshwork ablation with trabectome has a reported a 55% success rate at 1 year.64 Unpublished results of a stent from the anterior chamber to canal of Schlemm (Glaukos i-stent, Laguna Hills, CA, USA) inserted ab interno suggest a 66% (IOP ≤ 21) success at 18 months.64,67 Additionally, excimer laser trabeculotomy has been described applying laser ab interno with success rates of 41 to 91% at 1 year.68

In summary, current glaucoma surgical practice involves the judicious assessment of patients and the judgement of whether medical treatment is adequate. It is paramount to adequately measure glaucoma progression and clearly justify surgical intervention as there are significant risks. It is also paramount to fully assess the patient and determine the most effective and safest procedure best suited to their condition. The current mainstays of surgery, being trabeculectomy, GDDs and paediatric trabeculotomy and goniotomy, have been in existence for more than 40 years and have been refined somewhat. Currently, a lot of work is taking place to develop newer potentially safer forms of glaucoma surgery, but unfortunately, none have been shown to be superior to the traditional procedures.