Over the last 100 or so years, many surgical approaches to glaucoma have been proposed and tried. Few have had the staying power of trabeculectomy or, more recently, the tube to posterior reservoir procedures. These are relatively easy to perform, generally have good success rates and have only rare disastrous consequences. On the other hand, the results are not always predictable; serious, if not disastrous, complications occur with alarming frequency and, with the advent of antifibrotic agents, serious late complications, such as failure, bleb leaks, blebitis, and even endophthalmitis, are occurring at increasing and disturbing rates. In addition, recent large-scale, prospective studies have shown the desirability of long-term, relatively tight pressure control in advanced glaucoma. Finally, as the problems of long-term medical therapy such as poor adherence, cost, side effects, and interference with the quality of life become more apparent, a simple, safe, and successful surgical procedure would be most welcome, especially for those patients with poor adherence for whatever reasons and in those parts of the world where the costs of chronic medical therapy are prohibitive.
We actually have several good operations for glaucoma including trabeculectomy with and without adjunctive antifibrotic agents, anterior chamber to posterior equatorial reservoir tube shunts, and cyclo-photocoagulation. However, in addition to unpredictability the long-term results are disappointing. For trabeculectomy, the complications of choroidal hemorrhage, hypotony, early and late bleb leaks, and early and late endothalmitis can be daunting. For tube-shunt procedures, the need for supplemental antiglaucoma medications, late failure, and corneal decompensation are serious problems. Finally, for cyclophotocoagulation, the need for repeat procedures and the risks of uveitis, and phthisis bulbi are causes for concern.
Over the last decade or so, several new procedures have been described which seek to address some of the problems of these three general approaches. Procedures have been devised to bypass only the trabecular meshwork. These procedures shunt aqueous from the anterior chamber into Schlemm’s canal. Other procedures have been devised to improve flow within the canal by dilating it; this type of operation is called viscocanalostomy. Another procedure bypasses the intrascleral channels by unroofing Schlemm’s canal and creating an intrascleral reservoir. This latter procedure is called ‘non-penetrating’ deep sclerectomy. The Express shunt bypasses the entire drainage apparatus like a trabeculectomy into an anterior subTenon’s space using a stainless steel needle-like device with an internal flow restrictor. Finally, two devices are proposed for shunting aqueous from the anterior chamber into the suprachoroidal space. Each of these will be discussed in turn with the known results, pros and cons. It is well to keep in mind that glaucoma is usually a lifetime disease and that, for each new procedure, long follow-up and randomized clinical trials will be necessary to determine the real place of each in the surgical armamentarium. Several procedures that seemed to make sense and had early successes did not stand the test of time.
NON-PENETRATING GLAUCOMA SURGERY
Back in the 1950s Epstein observed aqueous oozing from the site of thin sclera resulting from the deep removal of pterygia. Krasnov actually described a procedure, called sinusotomy, to unroof Schlemm’s canal over 180° allowing aqueous to drain subconjunctivally. This early procedure was abandoned when it was found that the pressure-lowering effects were relatively short lived (and perhaps also because of the near simultaneous introduction of trabeculectomy). The concept was revived in the 1980s by Zimmerman and co-workers who described a procedure which was similar to sinusotomy but was performed under a scleral flap and which could be followed postoperatively, if necessary, with a yttrium-aluminum-garnet (YAG) laser goniopuncture. The theoretical advantages of non-penetrating surgery include: less chance of both early and late infection; reduced risk of sudden and prolonged hypotony; less bleeding, and less chance of either serous or hemorrhagic choroidal detachment. As a general rule, these theoretical advantages have proven correct in practice. Whether adequate pressure control over the long haul is attained is still being debated.
The two main types of non-penetrating glaucoma surgery are the viscocanalostomy, where the primary focus is on dilating Schlemm’s canal, and the deep sclerectomy, where the primary focus is on unroofing Schlemm’s canal (and often the inner wall thereof ) and creating an intrascleral reservoir with or without an intrascleral implant. Neither type of operation is as standardized as the trabeculectomy. There are overlapping features of both types of procedures and some surgical steps are used in both ( Boxes 38-1 and 38-2 ).
No sudden decompression of anterior chamber
Suprachoroidal hemorrhage less likely
Serous choroidal detachment less likely
Reduced risk of prolonged hypotony
Less likely to get filtering bleb
Less chance of bleb leak – early or late
Less chance of blebitis, endophthalmitis
Contact lens wear less likely to be problematic
Bleb dysthesia rare
Less intraocular inflammation
Less chance of intraocular bleeding
Fewer postoperative visits
More rapid visual rehabilitation postoperatively
Technically more difficult
Takes longer in the operating room
Requires some specialized instrumentation
About 10% have actual perforation into anterior chamber requiring iridectomy
Intraocular pressure less likely to be lowered sufficiently in advanced glaucoma
Pressure lowering may not last as long
Stegman contributed to the ‘new’ surgery trend with his description of a procedure in which Schlemm’s canal was enlarged by viscoelastic material injected through a fine canula from an external dissection unroofing the canal. In addition, a window in Descemet’s membrane was created to allow aqueous access to an intrascleral space. Theoretically, this procedure did not enter the anterior chamber but opened Schlemm’s canal and, in addition, created an intrascleral reservoir (by removing an internal scleral block) from which the aqueous drained via intrascleral vascular channels. As noted above, operations unroofing Schlemm’s canal without theoretically entering the anterior chamber had been previously described by Krasnov, Nesterov and Zimmerman in a limited number of patients but Stegman’s procedure was presented with over 200 cases in black South Africans with reasonably long follow-up and impressive results. The success rate was over 80% defined as intraocular pressure (IOP) less than 22 mmHg without need for supplemental antiglaucoma medications. The follow-up averaged almost 3 years.
Other groups had good results but not as good as Stegman. One group reported 36% success without medications and 73% with supplemental antiglaucoma medications at 1 year but with only a 2% complication rate. A few small randomized trials comparing primary viscocanalostomy against trabeculectomy, from Germany, England, Japan and Italy, showed lower IOPs with trabeculectomy and shorter duration of IOP lowering with viscocanalostomy but lower complication rates with viscocanalostomy. In one of these studies, complete success at 1 year (no medications) was found in 100% of the trabeculectomy patients but in only 64% of the eyes with viscocanalostomy. This same group found no serious complications over the follow-up period with either procedure in the 50 eyes. A randomized study of 50 eyes from Japan showed an average IOP at 1 year of 17 mmHg in the eyes having viscocanalostomy and 12 mmHg in eyes having trabeculectomy with mitomycin-C. The American Academy of Ophthalmology, through its technology assessment committee, found that of 100 citations in the literature, viscocanalostomy or deep sclerectomy succeeded in lowering IOP into the mid to upper teens most of the time with a low complication rate, especially as it relates to overfiltration and hypotony. The committee also noted the need for more randomized, controlled trials. One 3-year prospective study of 67 eyes noted a complete success rate (target IOP without medications) of 60% at 3 years and a total success rate of 88% without any long-term complications. Yet another prospective study of 57 eyes followed for 5 years after viscocanalostomy found IOP less than 21 mmHg without medications in 60% and with or without medications in 90% at 5 years. In this latter study, more than one-third of the eyes needed postoperative YAG laser goniopuncture to achieve adequate pressure lowering at an average time of 9 months after surgery. Intraocular pressure lowering after goniopuncture averaged 8 mmHg.
Gimbel and co-workers described combining this procedure with cataract extraction and lens implant. Wishart and co-workers reported on a 2-year follow-up of viscocanalostomy and phacoviscocanalostomy in 101 eyes; 93% had IOPs less than 21 mmHg at 2 years. Phacoemulsification combined with viscocanalostomy achieved lower IOPs than phacoemulsification alone, with a low complication rate in one non-randomized study. In a non-randomized trial, a group in Japan compared combined phacoemulsification and viscocanalostomy (phacoviscocanalostomy) with combined phacoemulsification and trabeculotomy in over 100 eyes with primary open-angle glaucoma; they found similar results in both groups.
The effectiveness of peeling of the juxtacanalicular tissue has been the subject of some debate; in one prospective study, it did not affect the long-term outcome but did reduce slightly the incidence of immediate postoperative pressure spikes. Johnson and Johnson attempted to determine the mechanism of action of viscocanalostomy and felt that it worked by making microperforations in the trabecular meshwork. Wild and co-workers were unable to show any perforations through trabecular meshwork in eye-bank eyes subjected to viscocanalostomy. Furthermore, they demonstrated dilation of Schlemm’s canal up to 6 mm from the injection site and a small increment in the length of dilation using a more viscous viscoelastic. In an extensive study of primate and human eyes, Smit and Johnstone demonstrated disruption of the structure of Schlemm’s canal, both inner and outer walls, as well as loss of bridging elements for as much as 16 mm from the site of injection. Furthermore, intrascleral channels were dilated. They concluded that viscocanalostomy may work by allowing direct access of aqueous from the juxtacanalicular tissue to the intrascleral collector channels. In Rhesus monkeys, viscocanalostomy appears to work by disrupting the inner wall of Schlemm’s canal and opening the juxtacanalicular region to easier aqueous access. A similar picture was obtained in two living human eyes using an endoscope during and after viscoelastic injection into Schlemm’s canal.
The presence of the intrascleral chamber and the absence of a filtering bleb characterized the ultrasound biomicroscopic (UBM) appearance of postoperative eyes with successful viscocanalostomy. In another study using the ultrasound biomicroscope in eyes following viscocanalostomy, perforation of the Descemet’s window and the presence of the intrascleral ‘lake’ were correlated with successful surgery. However, in yet another study, the intrascleral lake was not demonstrable on UBM in one-third of successful patients. A concave appearing trabeculo–Descemet’s membrane by gonioscopy or UBM heralded increased IOP in postoperative eyes with viscocanalostomy; the membrane became flat after Nd:YAG goniopuncture and the IOP was reduced.
Viscocanalostomy has also been shown to be effective in congenital glaucoma – at least as effective as trabeculotomy ab externo and perhaps more effective in more aggressive disease. The procedure also seems to have a high success rate (80%) in juvenile-onset glaucoma rivaling that reported for trabeculotomy. One small study suggests that viscocanalostomy may be a reasonable alternative to trabeculectomy in eyes with uveitic glaucoma.
Complications include failure to find Schlemm’s canal and perforation or microperforations into the anterior chamber requiring an iridectomy and/or conversion to traditional trabeculectomy. Dietlein and co-workers examined 177 trabeculectomy specimens and found that older patients, in particular, had thin trabecular meshwork suggesting that perforation may be more likely in this group. Hyphema and hypotony also can occur with consequent serous choroidal detachment, although less frequently than seen with trabeculectomy.5 Perforation of Descemet’s membrane and hyphema each occurred at a rate of 10%. Detachment of Descemet’s membrane requiring surgical repair may happen. However, these may spontaneously recover without intervention and so should be watched for some time before surgical repair is attempted. Injection of viscoelastic into the corneal stroma has also been reported as has one case of intracorneal hematoma. One case of delayed suprachoroidal hemorrhage has been published.
Several variations on the theme of viscocanalostomy have been proposed. Klink and colleagues reported the use of an erbium laser to assist in removing the inner scleral block. Lewis presented experience with a Schlemm’s canal catheter which can inject viscoelastic substance 360°; this same catheter can be used to thread a suture around the entire circumference of Schlemm’s canal and can lower IOP by exerting a little tension on the suture so it expands trabecular meshwork into the anterior chamber ( Fig. 38-1 ). Early studies with this technique by Stegman have been promising but long-term follow-up and larger numbers will be needed to determine if this variation on the technique will find a place in the surgical armamentarium.
In summary, viscocanalostomy is a promising new surgical approach that effectively lowers IOP; compared to trabeculectomy, it lowers IOP less well, is more difficult to perform, and takes longer, but complications, especially those related to hypotony, are fewer and less severe.
BYPASS INTRASCLERAL CHANNELS (NON-PENETRATING DEEP SCLERECTOMY)
Building on the descriptions and experience of Krasnov, Zimmerman and Stegman with their procedure, others in Europe and Japan described a procedure in which an internal scleral block was removed under an outer scleral flap, Schlemm’s canal was unroofed, and Descemet’s membrane was exposed under the outer flap. Demailly and co-workers added a collagen implant to prolong the presence of the internal scleral reservoir. Transgonioscopic neodymium (Nd):YAG laser goniopuncture could be added at a later date if the pressure started to rise. Mermoud and his group in Switzerland became fans of deep sclerectomy and were able to show that the collagen implant improved the outcome over deep sclerectomy without it. Ultrasound biomicroscopy 1 month after the deep sclerectomy with collagen implant (DSCI) showed filtration around the scleral flap into the subconjunctival space and possibly some suprachoroidal filtration ( Fig. 38-2 ). Another proposed mechanism of action of this procedure is to increase the intrascleral vessels collecting aqueous from the intrascleral reservoir. The collagen implant dissolved by about 9 months but the pressure-lowering activity continued ( Fig. 38-3 ).
The results of DSCI have been reasonable with a reduction in IOP that is clinically significant and with a significant reduction in complications. Average IOP at 1 year has run about 12 mmHg and, in a matched, case control series, success as defined by IOP less than 21 mmHg without adjunctive medication was 69% as compared to that with trabeculectomy at 57%. In this same group, complications with DSCI were fewer and less serious than those seen with trabeculectomy. In a series followed for up to 3 years, success using the above criterion was 44% and qualified success (controlled under 21 mmHg with topical medications) was 97%. In a randomized, prospective trial of deep sclerectomy versus trabeculectomy from Saudi Arabia, deep sclerectomy produced an average IOP reduction at 1 year of 12 mmHg compared to 14 mmHg for trabeculectomy (not statistically significantly different). Success as defined by IOP equal to or less than 21 mmHg was 92% for deep sclerectomy versus 95% for trabeculectomy, but the rate of flat or shallow chambers was 7% for trabeculectomy and 0% for deep sclerectomy. In yet another study of intermediate-term results, DSCI lowered IOP from a preoperative level of 25 mmHg to 15 mmHg postoperatively with only about one-half requiring adjunctive topical medication (mostly betaxolol). Similar results were obtained with a reticulated hyaluronic acid implant although a higher per cent of these patients seem to have a visible filtering bleb. Deep sclerectomy without implant lowers pressure less well than trabeculectomy in randomized, controlled trials, although complications do seem to be fewer than trabeculectomy with deep sclerectomy. In a study where one eye was given a deep sclerectomy with implant and the other eye no implant, the eyes with the collagen implants had the lower IOPs and required fewer adjunctive topical antiglaucoma medications with no difference in complication rate. In another randomized, prospective study by the same group, the collagen implant eyes had lower IOPs and a better success rate without medications than those with a deep sclerectomy and no implant. It seems reasonable to conclude that the implants do enhance the success rate of deep sclerectomy.
Shaarawy and co-authors reported on 105 eyes in 105 patients followed for 5 years. They found an average IOP of 11 mmHg with 95% of patients below 21 mmHg and 45% not requiring adjunctive medication for control. In another study by the same group in which trabeculectomy was performed in one eye and deep sclerectomy with collagen implant in the other, the results were comparable although pressures were slightly lower for the eyes having trabeculectomy: 12.9 mmHg at 2 years compared to 13.9 mmHg for the DSCI eyes with equal overall success rates; however, the DSCI eyes had 50% less hyphema and serous choroidal detachments in the early postoperative period. One almost 4-year study in patients with exfoliative glaucoma showed roughly half controlled at 19 mmHg or below without medications and only one-third of patients with primary open-angle glaucoma controlled by the same criteria; the authors conclude that deep sclerectomy with implant works better in eyes with pseudoexfoliation than in eyes with primary open-angle glaucoma. Over 2 years, deep sclerectomy without implants, mitomycin-C, or goniopuncture works almost as well as trabeculectomy without mitomycin-C, but the trend of IOP suggests some loss of efficacy over time, although results of DSCI may be maintained even up to 8 years of follow-up.
Postoperative Nd:YAG laser goniopuncture is required to control pressure in anywhere from 3% to 80% with the average settling around 50% with long-term follow-up ; intraoperative mitomycin-C application seems to reduce the need for this maneuver. In one study, IOP fell from a pre-laser average level of 32 mmHg to an immediate post-laser level of 17 mmHg with maintenance of this level for at least 6 months. Complications of Nd:YAG goniopuncture have been limited to iris incarceration into the site with subsequent elevation of IOP.
Filtering blebs have been small or not observed; inflammation is reduced compared to trabeculectomy perhaps because no iridectomy is performed.
Despite the fact that this operation, theoretically at least, does not depend on filtration under the conjunctiva for success, the success rate appears enhanced with the use of intraoperative application of mitomycin-C to the operative site. However, in one randomized clinical trial in west African patients, the authors were unable to show any difference up to 18 months between those having deep sclerectomy with mitomycin-C and those not receiving mitomycin. In this same study, while the results at 1 year were mildly encouraging with about 70+% having IOPs under 18 mmHg, the results at 18 months were quite disappointing in both groups with success hovering only around 35%.
As with trabeculectomy, the operation seems to work better in eyes not previously treated with topical medications for glaucoma. The procedure also seems to work in highly myopic eyes despite the thin sclera. Success in eyes with uveitic glaucoma has also been reported.
Deep sclerectomy with collagen implant can be successfully combined with phacoemulsification cataract surgery. In one retrospective study of consecutive cases, there was no difference in pressure control or major complications between combined phacoemulsification with trabeculectomy and phacoemulsification with deep sclerectomy except a significantly higher risk of bleb leaks in the phacoemulsification with trabeculectomy group.
Complications include perforation of the trabecular meshwork with need to convert to trabeculectomy, scleral ectasia, iris incarceration, hemorrhagic Descemet’s detachment, hypotony, and vitreous hemorrhage. Bleeding into the anterior chamber from the site of deep sclerectomy during gonioscopic examination can occur quite late in the postoperative period ( Box 38-3 ). Having to convert to trabeculectomy because of perforation into the anterior chamber through Descemet’s membrane or trabecular meshwork produces, as expected, ultimately a lower IOP than uncomplicated deep sclerectomy but increases the early postoperative complication rate and prolongs the time to recovery of best vision.