Glaucoma Drainage Devices 

Glaucoma drainage devices (GDDs) create an alternate aqueous pathway from the anterior chamber (AC) by channeling aqueous out of the eye through a tube to a subconjunctival bleb or to the suprachoroidal space. This tube is usually connected to an equatorial plate under the conjunctiva. GDDs are being used more frequently in the treatment of glaucoma that is not responding to medications and trabeculectomy operations. In certain conditions, such as neovascular glaucoma, iridocorneal endothelial (ICE) syndrome, penetrating keratoplasty (PKP) with glaucoma, and glaucoma following retinal detachment surgery, it has become the preferred operation. This article outlines the current concepts involving different GDDs, surgical techniques, and management of complications following GDD insertion.

The earliest attempt to drain fluid out of the anterior chamber into the subconjunctival space at the limbus dates back to 1906 when Rollet and Moreau implanted a silk thread connecting the anterior chamber to the subconjunctival space. Since that time, additional unsuccessful attempts were made, including insertion of a polythene tube by Epstein in 1959 and a silicone tube by MacDonald and Pearce in 1965. These operations failed because of excessive scar formation near the limbus, seton migration, and conjunctival erosion.

In 1969, Molteno introduced the concept that a large surface area was needed to disperse the aqueous beneath the conjunctiva. He inserted a short acrylic tube that was attached to a thin acrylic plate. The plate was sutured to the sclera close to the limbus. Most of the operations failed after the first 3-6 months because of plate exposure, tube erosion, and scar formation.

In 1973, Molteno improved his device with the idea of draining the fluid away from the limbus to increase the success rate. He introduced the Molteno implant with a long silicone tube attached to a large end plate placed 9-10 mm posterior to the limbus.^[1] All the currently available GDDs are based on this concept by Molteno. The Molteno implant and similar nonvalved implants offer no resistance to the outflow, resulting in hypotony, flat anterior chambers, and choroidal effusions.

Since then, 2 major concepts have been introduced to modify the implantation of the GDD.

The first approach was that of a valve to offer resistance to the outflow, thereby reducing the incidence of postoperative hypotony. In 1976, Krupin developed a pressure-sensitive, unidirectional valve that provides resistance to the flow of aqueous and prevents early postoperative hypotony. This "slit valve" is designed to open at a pressure of 11 mm Hg and to close at a pressure of 9 mm Hg.

In 1993, Ahmed introduced the Ahmed glaucoma valve (AGV), a pressure-sensitive, unidirectional valve that is designed to open when the intraocular pressure (IOP) is 8 mm Hg (see image below).^{[2, 3, 4]}

The second major change has been the realization that by increasing the surface area of the end plate, the surface area of drainage could be increased, resulting in lower IOPs.^{[5, 6, 7]}

In 1981, Molteno introduced the double plate implant with a surface area of 270 mm². In 1992, Baerveldt introduced a nonvalved silicone tube attached to a large barium-impregnated silicone plate with a surface area of 250 mm², 350 mm², or 500 mm² (see image below).^{[8, 9, 6, 10, 11]}

Optonol Ltd developed the Ex-PRESS R50 glaucoma shunt to simplify the GDD implantation. This device is a single-piece, stainless steel, translimbal implant that is placed using an inserter. Although its implantation is efficient, the long-term efficacy and the risk of complications have yet to be determined.

Current glaucoma drainage devices

Current GDDs can be classified into those with no resistance, those with resistance, and those with variable resistance to aqueous outflow.

GDDs with no resistance

These GDDs consist of a silicone tube attached to an end plate that acts as a surface for bleb formation. Unless the operation is modified with a stent and ripcord technique, these implants are associated (in the early postoperative period) with a high incidence of overfiltration secondary to no aqueous outflow resistance. This can lead to hypotony, shallow-to-flat anterior chambers, and choroidal effusions.

• The single-plate Molteno implant is a silicone tube attached to a 135 mm² polypropylene end plate.
• The double-plate Molteno (DPM) is the same as the single-plate Molteno except that a second end plate is attached to the right or left of the original end plate, thus doubling its surface area. It requires a 2-quadrant dissection.
• The Baerveldt implant was developed to provide easy placement of a large end plate in a single quadrant. It is a silicone tube attached to a soft, pliable, barium-impregnated silicone end plate of various sizes (ie, 200 mm², 250 mm², 350 mm², 500 mm²). The placement of the end plate wings underneath the rectus muscles can promote fibrous encapsulation, resulting in disturbing diplopia. The design has been modified with fenestrations in the end plate that may allow fibrous tissue tacks to limit bleb elevation. Although not typical, diplopia continues to be a risk with this implant.
• The Schocket implant (anterior chamber tube shunt to encircling band [ACTSEB]) consists of a silastic tube used for nasolacrimal intubation. One end of the implant is inserted into the anterior chamber, while the other end is tucked underneath a No. 20 retinal-encircling band placed underneath the rectus muscles. Even though the procedure is lengthy and cumbersome, it is less expensive, and material can be assembled in most operating rooms.
• The Ex-PRESS R50 implant is a 3-mm long tube with a 400 µm (27 gauge) external diameter and a 50 µm internal diameter. The penetrating tip is beveled with 3 side orifices and a spurlike projection to prevent extrusion. The flange is a small (< 1 mm²) disclike plate that prevents the device from being inserted too deeply. Both the flange and the spurlike projection are angled to conform to the sclera with the distance between them being equal to the scleral thickness at the site of proper implantation.

GDDs with set resistance

Even though manufacturers claim that these devices contain true valves, independent examinations of the flow characteristics for these devices suggest a wide divergence between observed function and the manufacturers' claims. The valves appear not to close after initial opening in perfusion tests at physiological flow rates. Of the 2 valved devices that are used commonly, the AGV has the lowest incidence of hypotony of all GDDs.

The AGV is a silicone tube connected to a silicone sheet valve held in a polypropylene body. The end plate measures 185 mm² (16 mm long X 13 mm wide X 1.9 mm thick). The valve consists of thin silicone elastomer membranes (8 mm long X 7 mm wide) that create a venturi-shaped chamber. The inlet cross-section of the chamber is wider than the outlet (Bernoulli principle), with a resultant pressure differential between the anterior chamber and the bleb. The valve is designed to open when the IOP is 8 mm Hg.

The Krupin slit valve consists of a silicone tube with a slit valve attached to a silicone oval end plate. The surface area of the end plate is 180 mm². The opening pressure of the slit valve is designed to be 11-14 mm Hg, and the closing pressure is designed to be 2 mm Hg. Unfortunately, these opening and closing pressures may vary significantly.

GDDs with variable resistance

These devices may be modifications of the original Molteno implant and the Baerveldt implant that attempt to incorporate a resistance mechanism dependent on tissue apposition to limit flow. Because the force of tissue apposition is variable, these devices do not function as true valves, and IOP levels remain unpredictable.

The Molteno dual ridge device (Molteno with a pressure ridge) attempts to limit the initial drainage area by dividing the top portion of the plate into 2 separate spaces with the help of a thin V-shaped ridge. Aqueous must overcome the overlying conjunctival resistance to flow across the ridge. The resistance offered by the overlying conjunctiva presumably prevents overfiltration and hypotony. In the authors' experience, these complications are not prevented by the pressure ridge mechanism, so the authors still recommend a stent with ripcord modification.

The Baerveldt bioseal is a flap that overhangs the silicone tube as it opens on the end plate. Apposition of the bioseal element to the sclera with absorbable sutures is supposed to provide early flow resistance, limiting initial aqueous escape from beneath the device. However, early clinical trials failed to prove this concept, and this modification was discontinued.

The SOLX Gold Shunt is an investigational device consisting of a flat, 24-carat gold implant (5.2 mm long and 3.2 mm wide) with numerous microtubular channels that bridge the anterior chamber and the suprachoroidal space.^[12] The theoretical mechanism of the device is to shunt aqueous humor from the anterior chamber to the suprachoroidal space, similar to a cyclodialysis. It is an attempt to have a controlled cyclodialysis cleft. The initial design assumed that approximately half of the microchannels would be open initially, and the remainder could be opened after implantation using a 790-nm laser when a lower IOP is desired.

See image below for illustration of variable resistance devices.

Glaucoma incisional surgery is usually performed to establish definitive IOP control when medical therapy and laser surgery fail or are unable to be performed. Historically, GDDs have been reserved for patients who failed or were likely to fail trabeculectomy surgery. Usually, these high-risk eyes do not do well; thus, the GDD is recommended as a last resort procedure. However, in spite of the high-risk profile of patients enrolled into previous GDD studies, moderately good success with various designs of GDDs has been observed.^[13]

Two articles by Gedde and associates^{[14, 15]} provide the first multicenter, controlled clinical trial examining the efficacy and the outcomes of nonvalved GDDs versus trabeculectomy with mitomycin-C in similar patient populations with previous ocular surgery.

The first-year data provide evidence that, if confirmed with longer follow-up, will provide an evidence-based approach to the surgical management of complicated glaucoma.

The most provocative data presented in this study are the equivalent primary outcomes at 1 year in both the tube group and the trabeculectomy group. The mean IOP was not significantly different between the two groups. The trabeculectomy group had a larger percentage of complete successes (no adjunctive medical therapy), but the percentage of the overall success rate (eyes with or without supplemental medical therapy) was higher in the tube group. Failure was defined as IOP persistently greater than 21 mm Hg or not reduced by 20% from baseline, IOP less than 5 mm Hg, reoperation for glaucoma, or loss of light-perception vision. There was a higher failure rate in the trabeculectomy group (13.5%) than the tube group (3.9%) at 1 year. Certainly, the low rate of tube complications is encouraging for those who advocate an earlier and wider use of GDD surgery in glaucoma.

The indications for GDD implantation include the following:

• Neovascular glaucoma
• PKP with glaucoma
• Retinal detachment surgery with glaucoma
• ICE syndrome
• Traumatic glaucoma
• Uveitic glaucoma
• Open-angle glaucoma with failed trabeculectomy
• Epithelial downgrowth
• Refractory infantile glaucoma
• Contact lens wearers who need glaucoma filtration surgery

GDDs may have a complicated postoperative course. Thus, it is relatively contraindicated in patients unable to comply with self-care in the postoperative period. Borderline corneal endothelial function is a relative contraindication for anterior chamber placement of a tube.