As mentioned previously, the GDDs were traditionally indicated for the treatment of refractory glaucoma, meaning cases in which filter surgery is associated with a high risk of failure (for example, neovascular or uveitic glaucoma), with a high rate of complications or in cases with previous failure of filter surgery for glaucoma. Moreover, the shunts appear to be effective in patients in which a previous intraocular surgery (vitreo-retinal or corneal) led to the formation of conjunctival scar tissues that reduce the amount of intact conjunctiva, precluding the trabeculectomy procedure or function. In recent years, the indications for these devices have expanded quite considerably, and now include congenital/juvenile glaucoma, traumatic glaucoma, glaucoma in the aphakic/pseudophakic eye, glaucoma post-keratoplasty, other secondary glaucomas (Irido-Corneal-Endothelial syndrome, Epithelial downgrowth). Moreover, in eyes with residual visual function, the shunts may be preferred over the cyclodestructive procedures that run a high risk of blindness and Phthisis bulbi (end-stage eye) (see Table 5.2). Lastly, in recent years these devices have been proposed as the elective procedure for cases of primary open-angle glaucoma: a number of authors have suggested their use in the primary surgical management of non-complicated glaucoma—the most precocious cases—thanks to their better result predictability compared to the trabeculectomy with MM-C and the lower incidence of complications (see results and complications below).

It is essential that the surgeon performs a meticulous pre-operative evaluation of the eye and devises an appropriate surgical plan to ensure the good outcome of the implant and minimize the risk of complications. First of all, the mobility of the conjunctiva must be examined to allow the surgeon to choose the best quadrant for the implant. In children affected by buphthalmos or in patients with vascular collagen disorders, the surgeon should attempt to identify thinned areas of the sclera where the implant should be avoided: under these clinical conditions, the surgeon must use certain techniques to anchor the plate of the device to the sclera to avoid possible perforations. The surgeon should examine the patient’s cornea for marked peripheral gerontoxons or leukomas that could obstruct the visualization of the tube and cause problems with its insertion and correct positioning in the AC. Moreover, in the event of damage to the corneal endothelium, the implant in pars plana is preferable to the implant in AC. If the endothelial damage is evident (corneal insufficiency), and a clinically important cataract is observed, a triple procedure may be considered (phaco + IOL + GDD). The iris should be examined under high magnification to identify new blood vessels that may be treated with a preoperative intravitreal injection of anti-VEGF drugs to reduce the risk of intra- and post-operative bleeding. The depth of the AC must be measured to avoid exclude lesions to the cornea or the iris induced by the tube in the AC. The condition of the lens must also be considered: the tube can be positioned in the ciliary sulcus of pseudophakic eyes or in pars plana in aphakic eyes. It is extremely important that gonioscopy is performed on patients who are candidates for the GDD implant; this will identify neovessels in the camerular angle (in neovascular glaucoma) and peripheral anterior synechias (frequently present in neovascular, uveitic and traumatic glaucoma). The surgeon must take note of the areas that are free from anterior peripheral synechias as this will allow him to correctly choose the site for the implant. In the event of anterior peripheral synechias of recent appearance, the tube can be positioned anterior to the synechias. If synechias are observed in an extremely anterior position, it may be necessary to perform an intraoperative iridectomy to facilitate the insertion of the tube. Alternately, the surgeon must plan the implant of the tube in the ciliary sulcus or in pars plana.

In most cases, the tube of the GDDs will be positioned in the AC. However, it can also be positioned in the ciliary sulcus of a pseudophakic patient or in pars plana in vitrectomized eyes. The implantation of GDDs in AC will be examined more in detail later. Each surgical step of this procedure requires maximum attention to achieve good results and reduce the post-operative complications. The first thing to do is to select the quadrant for the implant. Most of the GDDs are positioned in a single quadrant, with the exception of the implants with two plates. When possible (due to the conditions of the conjunctiva, sclera, etc), the implant with a single plate should be positioned in the supero-temporal quadrant: this site consents the easiest access to implant the plate and leads to a lower rate of disturbance of ocular motility (see the paragraph on the complications below). With a silicone oil tamponade added to the eye, the implant is positioned in the temporal-inferior quadrant to prevent oil escaping; it is lighter than aqueous humor and will tend to rise.

When non-valved implants are used, two surgical maneuvers can be added to avoid hypotonia in the precocious post-operative period: both involve the use of suture thread that can be applied either inside or outside the tube of the implant.

The first of these additional maneuvers involves positioning an obstruction inside the tube (stent): it consists of a prolene or 3.0, 4.0 or 5.0 nylon thread, positioned inside the tube’s lumen. This will reduce the patency and consequently decrease the drainage of the aqueous humor: the thread acts as a temporary valve. The piece of thread is approximately 15–20 mm long; it is inserted in the tube for 6–10 mm through an opening on the tube close to the plate. The free end of the prolene or nylon thread that exits the tube is positioned in the subconjunctival space of the quadrant adjacent to the implant, close to the limbus. Once the fibrous capsule has formed around the plate (4–6 weeks from surgery), the suture can be removed under the slit lamp and topical anesthesia. At this point, the aqueous humor can drain freely into the fibrous capsule around the plate, with the capsule providing resistance to the flow. However, even though this technique is efficacious for preventing precocious postoperative hypotonia, it may be necessary to remove the suture immediately in the event of precocious postoperative hypertonia caused by occlusion of the tube. So there is again a risk of hypotonia, a reduction in the depth of the AC and the associated complications.

Most of the published studies report a mean percentage of success with the four types of implant of approximately 70% (range 50–80%), with an average postoperative reduction in the IOP of at least 50% over the preoperative values 2 years from surgery: the variation of the range depends on the type of glaucoma and the type of implant chosen. Unfortunately, failures account for 10% per year, with a consequent 50% function of the drainage implant after 5 years. One-year results of a comparative study for the Ahmed and Baerveldt implants did not clearly demonstrate any clear superiority of one over the other: even though the mean IOP was slightly higher in the eyes with the valved implant (Ahmed), this implant was associated with a lower incidence of precocious complications that are less severe than those observed with the Baerveldt implants. Recently an update with 5-year follow-up was published. It referred to an important study called ‘Treatment Outcomes in the Tube Versus Trabeculectomy” (TVT). It reported that 5 years from surgery, IOP was a slightly greater (though not significant) when shunts were used compared to trabeculectomy: in the postoperative, both types of procedure are associated with a fairly similar reduction in the IOP and comparable use of a topical pressure-reducing therapy. GDDs have a greater success rate compared to the trabeculectomy procedure associated with MM-C: the cumulative probability of failure during 5 years of follow-up has been reported as 29.8% for the GDDs and 46.9% for the trabeculectomy. By the same measure, the percentage of repeating surgery is greater for trabeculectomy. These important findings have driven numerous surgeons to increase the number of implant procedures performed and to propose this surgery in cases of simple, not complicated, chronic glaucoma. Barton et al concluded their study by stating that the Ahmed valves would appear to improve the predictability of the precocious postoperative control, and that the Baerveldt valves have a lower percentage of late encapsulation. The authors believe that the main obstruction to the diffusion of these devices in uncomplicated cases of glaucoma is the lack of data available on the long-term effects on the endothelium, an issue that has not been fully clarified. In conclusion, GDDs would appear to be associated with an efficacy that is comparable, and sometimes superior, to the trabeculectomy in the management of complicated glaucoma. Given that, traditionally, these devices were suggested for refractory cases of glaucoma, they have always been associated with poor results. Their use should be reviewed—something that is already happening to some degree—in the primary surgical management of uncomplicated glaucoma (the more precocious cases), even taking in consideration the greater predictability compared to the trabeculectomy with MM-C, combined with a lower incidence of complications.

There are numerous possible complications associated with GDDs; the surgeon must necessarily construct a precise preoperative plan and have learned a meticulous technique (for all phases of the procedure) to achieve a valid result.