Experimental Eye Research 93 (2011) 284e290

Contents lists available at ScienceDirect

Experimental Eye Research

journal homepage: www.elsevier.com/locate/yexer

Update on aqueous shunts

Steven J. Gedde a, Richard K. Parrish II a, Donald L. Budenz a, Dale K. Heuer b,* a Bascom Palmer Eye Institute, University of Miami, Miller School of Medicine, Miami, FL, USA b Department of Ophthalmology, Medical College of Wisconsin, Milwaukee, WI, USA article info abstract

Article history: Medicare claims data and surveys of the American Society membership show that aqueous Received 30 November 2010 shunts are being increasingly utilized in the surgical management of glaucoma. New clinical trials data Accepted in revised form 21 March 2011 have identified differences in the efficacy and safety of shunts in common use. Recent studies have Available online 31 March 2011 reported comparable results with and aqueous shunts in similar patient groups. Intra- operative and postoperative complications may develop with aqueous shunt surgery related to the Keywords: implantation of a foreign material. Several modifications in surgical technique have been directed toward aqueous shunts improving surgical success, reducing complications, and optimizing efficiency and cost. glaucoma drainage devices Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction capsule around the equatorial plate (Minckler et al., 1987; Wilcox et al., 1994). Therefore, the final IOP that is achieved after Glaucoma is a disease involving impaired outflow of aqueous aqueous shunt surgery is determined by capsular permeability and humor through the anterior chamber angle. The increased resis- surface area (i.e. more permeable and larger surface-area capsules tance to aqueous outflow produces (IOP) are associated with lower postoperative pressures). elevation that damages the optic nerve. The current treatment of Molteno pioneered the concept of wound-healing modulation glaucoma is directed toward decreasing IOP to prevent progressive to minimize the fibrous encapsulation of the equatorial explants in glaucomatous optic nerve damage. Glaucoma surgery is typically an effort to improve IOP control. He has advocated the combination performed when adequate IOP reduction cannot be achieved with of oral prednisone (or intramuscular methylprednisolone acetate), maximally tolerated medical therapy and appropriate laser treat- colchicine, and fluphenamic acid (a non-steroidal anti-inflamma- ment. Trabeculectomy and aqueous shunt surgery are the most tory agent) with topical atropine sulfate, epinephrine hydrochlo- commonly performed incisional glaucoma procedures worldwide. ride, and dexamethasone (or betamethasone sodium phosphate), Both operations create an alternative route for aqueous outflow, particularly for patients aged 18 months to 50 or 60 years in whom bypassing the high resistance pathway and providing IOP he regards the healing response most vigorous (Molteno et al., lowering. Trabeculectomy creates a scleral fistula that connects the 1976; Molteno, 1980, 1987). However, because others have experi- anterior chamber and subconjunctival space resulting in the enced or have been concerned about systemic side-effects, his formation of a paralimbal filtering bleb. Aqueous shunts are antifibrosis regimen has not achieved widespread application implantable devices that allow drainage of aqueous humor more (Brown and Cairns, 1983; Cairns, 1983). More recently, surgeons posteriorly. have attempted to minimize fibrovascular proliferation and thereby Aqueous shunts share a common design consisting of a silicone capsule thickness over the equatorial plate by applying mitomycin tube that is inserted into the eye through a scleral fistula and shunts C (MMC), but this approach appears to be of limited benefit(Cantor aqueous humor to an episcleral plate that is located in the equa- et al., 1998; Costa et al., 2004). Shunts with larger equatorial plates torial region of the globe, typically centered between (and, in the produce greater IOP reduction, presumably by promoting a larger case of some devices, extending under or over) two adjacent rectus surface-area capsule around the plate for filtration (Heuer et al., muscles. Fibrous encapsulation of the equatorial plate produces 1992). However, there appears to be an upper limit beyond which a reservoir into which aqueous humor pools. The major resistance an increase in plate size does not allow greater efficacy (Lloyd to aqueous outflow through these devices occurs across the fibrous et al., 1994). The equatorial plates of commercially-available aqueous shunts differ in the size, shape, and physical material composition. The * Corresponding author. material and surface characteristics of the equatorial plates may E-mail address: [email protected] (D.K. Heuer). also in fluence the wound-healing response elicited by these

0014-4835/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.exer.2011.03.013 S.J. Gedde et al. / Experimental Eye Research 93 (2011) 284e290 285 devices. In rabbits, subconjunctival implantation of polypropylene 2.2. Surveys of the American Glaucoma Society (AGS) induced more inflammatory reaction than silicone (Ayyala et al., 1999, 2000). Another recent study evaluated the surface topog- In late 1995, an anonymous survey of the AGS and Japanese raphy of four commercially-available aqueous shunts and found Glaucoma Society memberships was conducted by Chen et al. a correlation in tissue culture to Tenon’s capsule fibroblast adhesion (1997) to assess patterns of use of antifibrotic agents and aqueous to surface roughness as measured by three-dimensional white- shunts among glaucoma specialists. The survey presented ten light confocal microscopy (Choritz et al., 2010). clinical situations requiring glaucoma surgical intervention. For Aqueous shunts are “valved” or “non-valved”, depending on each clinical scenario, respondents were asked to estimate the whether a flow restriction mechanism limits aqueous humor flow. percentage in which they would perform trabeculectomy alone, Non-valved shunts require a temporary restriction of flow by tube trabeculectomy with 5-fluorouracil (5-FU), trabeculectomy with ligation or occlusion during surgical implantation. This maneuver MMC, or aqueous shunt implantation. Antifibrotic agents like 5-FU allows a fibrovascular capsule to develop around the equatorial and MMC are routinely used as adjuncts to glaucoma filtering plate before aqueous humor outflow begins and reduces the risk of surgery to reduce scarring at the surgical site and increase the hypotony in the early postoperative period. Valved shunts have likelihood of surgical success. The majority of respondents flow restrictors that are designed to lessen the likelihood of preferred trabeculectomy with MMC for each of the clinical hypotony before plate encapsulation. Valved shunts do not require scenarios. In 2002 Joshi et al. (2005) redistributed the same survey temporary flow restriction and offer the potential advantage of to AGS members. Respondents still favored trabeculectomy with immediate postoperative IOP reduction. Table 1 reviews the MMC for most clinical situations, but the percentage usage of advantages and disadvantages of commercially-available aqueous aqueous shunts substantially increased; specifically, selection of shunts (Schwartz et al., 2006). aqueous shunts averaged 17.5% in 1995 and 29.4% in 2002.

3. Surgical results with different implants 2. Surgical trends Glaucoma surgeons have questioned the superiority of the Medicare data and surveys of glaucoma specialists demonstrate commercially-available devices with respect to safety and efficacy. an increase in the utilization of aqueous shunts and a concurrent Several retrospective studies failed to detect a statistically signifi- decline in the popularity of trabeculectomy in recent years (Chen cant difference of surgical outcomes among different aqueous et al., 1997; Joshi et al., 2005; Ramulu et al., 2007). Concern about shunt types (Ayyala et al., 2002; Nassiri et al., 2010; Smith et al., bleb-related complications, such as bleb leaks, blebitis, and bleb- 1995; Syed et al., 2004; Tsai et al., 2003, 2006; Wang et al., 2004). related endophthalmitis (DeBry et al., 2002), has likely contributed The conclusions of these studies are limited by inherent biases of to the expanded use of aqueous shunts as an alternative to trabe- retrospective investigations. Recent prospective clinical trials have culectomy. Aqueous shunts were initially reserved for the treat- provided higher level medical evidence regarding popular ment of eyes with that had poor surgical prognoses with currently utilized aqueous shunts (Barton et al., 2011; Budenz et al., standard filtering surgery even with wound-healing modulation. 2011; Nassiri et al., 2010). Table 2 provides a summary of key However, an increasingly positive experience with these devices elements of these randomized clinical trials. has prompted their implantation in glaucomas with better surgical prognoses (Jamil and Mills, 2007). 3.1. Ahmed Glaucoma valve implant vs Molteno implant

2.1. Medicare utilization data Nassiri et al. (2010) reported the results of a prospective randomized study that compared the Ahmed glaucoma valve Ramulu et al. (2007) examined trends in utilization of the implant (New World Medical, Inc., Rancho Cucamonga, California, most common glaucoma-related laser and surgical treatments USA) and single-plate Molteno implant (Molteno Ophthalmic for Medicare beneficiaries over the period 1995e2004. The Limited, Dunedin, New Zealand) in 92 patients with refractory numbers of procedures performed annually were tabulated with glaucoma. The rate of surgical failure (IOP > 21 mm Hg, IOP 5mm Current Procedural Terminology (CPT) codes. Trabeculectomies Hg, phthisis bulbi, loss of light perception vision, removal of the decreased by 43% and placement of aqueous shunts increased by implant, reoperation for glaucoma, or any devastating intra- 184% between 1995 and 2004. Commenting on those trends, operative or postoperative complication) was similar for both Corcoran (2009) noted that in 1995, that there had been 23 treatment groups after 2-year follow-up (16% Molteno group vs 18% times as many trabeculectomies as aqueous shunts, but by Ahmed group). The Molteno group had a greater percentage drop in 2007 that ratio had changed to only three times more IOP from baseline at 2 years (49.7% Molteno group vs 41.9% Ahmed trabeculectomies. group, p ¼ 0.049), but the mean number of glaucoma medications

Table 1 Advantages and disadvantages of commercially-available aqueous shunts.

Shunt Type Commercially-available shunts Advantages Disadvantages Non-valved Baerveldt glaucoma Larger surface area of end plate(s) Delayed functioning until encapsulation of plate occurs implant, Molteno implant (Baerveldt glaucoma implant and Greater risk of hypotony-related complications Molteno double-plate implant) provides greater surface area that may increase IOP-reducing efficacy

Valved Ahmed glaucoma valve Allows immediate IOP reduction Higher rate of bleb encapsulation Valve minimizes risk of hypotony-related Smaller surface area of end plate may decrease complications IOP-reducing efficacy Ease of implantation Valve malfunction can result in hypotony or obstructed outflow 286 S.J. Gedde et al. / Experimental Eye Research 93 (2011) 284e290

Table 2 Summary of recent randomized clinical trials involving aqueous shunts.

Clinical trial Number of Treatment groups Follow-up Major findings patients/sites Ahmed Baerveldt 276/16 Ahmed glaucoma valve 1 year Mean IOP was lower with Baerveldt implant (15.4 mm Hg Ahmed vs 13.2 mm Comparison FP7, 350-mm2 Baerveldt Hg Baerveldt) (ABC) study glaucoma implant Use of glaucoma medications was similar between implants (1.8 Ahmed vs 1.5 Baerveldt) No significant difference in failure rates between implants (16.4% Ahmed vs 14.0% Baerveldt) Ahmed implant had a lower rate of early postoperative complications (43% Ahmed vs 58% Baerveldt) and serious complications (20% Ahmed vs 34% Baerveldt) Rate of late postoperative complications was similar between implants (29% Ahmed vs 37% Baerveldt)

Ahmed and 92/3 Ahmed glaucoma valve 2 years Similar failure rates with both implants (16% Molteno vs 18% Ahmed) single-plate Molteno FP7, Single-plate Molteno implant produced a greater percentage drop in IOP from baseline (49.7% implant study Molteno implant Molteno vs 41.9% Ahmed) No significant difference in mean glaucoma medications between implants (1.41 Molteno vs 1.03 Ahmed)

Ahmed implant vs 123/1 Ahmed glaucoma valve S-2, Mean Similar mean IOP (13.1 mm Hg Ahmed vs 13.6 mm Hg trabeculectomy) and mean trabeculectomy study Trabeculectomy 31 months number of glaucoma medications (1.13 Ahmed vs 0.93 trabeculectomy) with both surgical procedures No significant difference in success rates between procedures (69.8% Ahmed vs 68.1% trabeculectomy)

Tube Versus 212/17 350-mm2 Baerveldt glaucoma 3 years Aqueous shunt surgery had a higher success rate (84.9% tube vs 69.3% Trabeculectomy implant, Trabeculectomy with trabeculectomy) (TVT) study mitomycin C Similar mean IOP (13.0 mm Hg tube vs 13.3 mm Hg trabeculectomy) and mean number of glaucoma medications (1.3 tube, 1.0 trabeculectomy) with both surgical procedures Trabeculectomy had a higher rate of postoperative complications (39% tube vs 60% trabeculectomy) No significant difference in the rate of serious complications (22% tube vs 27% trabeculectomy)

Budenz et al. (2011), Gedde et al. (2009), Nassiri et al. (2010), Wilson et al. (2003). was not significantly different between the two groups (1.41 Mol- 37% Baerveldt group, p ¼ 0.16). The larger equatorial plate of the teno group vs 1.03 Ahmed group). Patients were censored from Baerveldt implant likely explains the lower IOP relative to the analysis at the time of failure, which was related to inadequate IOP Ahmed implant, as larger surface-area plates are associated with reduction in all patients. Postoperative complications were greater pressure reduction (Heuer et al., 1992). The lower incidence comparable between the Molteno and the Ahmed groups, and no of early postoperative and serious complications after placement devastating complications were observed in either treatment group. of the Ahmed implant compared with the Baerveldt implant probably relates to the valve mechanism that minimizes the risk 3.2. Ahmed Baerveldt Comparison (ABC) study of hypotony-related complications in the immediate postoperative period. The ABC Study is a multicenter randomized clinical trial that was designed to compare the efficacy and safety of the Ahmed glau- 4. Aqueous shunts versus trabeculectomy coma valve implant and 350-mm2 Baerveldt glaucoma implant (Abbott Medical Optics, Abbott Park, Illinois, USA) in 276 patients Despite the introduction of several new incisional glaucoma with refractory glaucoma (Barton et al., 2011). Mean IOP was lower procedures in recent years, trabeculectomy and aqueous shunt after Baerveldt placement at 1 year (15.4 mm Hg Ahmed group vs surgery remain the most frequently performed glaucoma opera- 13.2 mm Hg Baerveldt group, p ¼ 0.007), and use of adjunctive tions worldwide. No clear consensus exists among glaucoma glaucoma medications was similar between treatment groups (1.8 specialists regarding the best surgical approach for managing medications Ahmed group vs 1.5 medications Baerveldt group, medically uncontrolled glaucoma in many clinical settings. Some p ¼ 0.07) (Budenz et al., 2011). The cumulative probability of failure surgeons favor an antifibrotic-augmented trabeculectomy, while (IOP > 21 mm Hg or not reduced 20% from baseline, IOP 5mm others prefer the placement of an aqueous shunt (Chen et al., 1997; Hg, additional glaucoma surgery, removal of the shunt, or loss of Joshi et al., 2005). Several studies have provided information light perception vision) was not significantly different between the comparing surgical results with aqueous shunts and trabeculec- two groups after 1-year follow-up (16.4% Ahmed group vs 14.0% tomy in specific patient groups (Gedde et al., 2005, 2007a, 2007b, Baerveldt group, p ¼ 0.52). More patients who received a Baerveldt 2009; Stein et al., 2008; Tran et al., 2009; Wilson et al., 2000, implant experienced early postoperative complications during the 2003). Table 2 offers a summary of key elements of randomized first 3 months after surgery (43% Ahmed group vs 58% Baerveldt clinical trials comparing aqueous shunts and trabeculectomy. group, p ¼ 0.016) and serious complications associated with reoperation and/or vision loss of two or more Snellen lines 4.1. Surgical outcomes among medicare beneficiaries (20% Ahmed group vs 34% Baerveldt group, p ¼ 0.014). No signifi- cant difference in the rate of late postoperative complications was Stein et al. (2008) assessed the rates of postoperative adverse observed between the two treatment groups (29% Ahmed group vs outcomes after trabeculectomy and aqueous shunt procedures S.J. Gedde et al. / Experimental Eye Research 93 (2011) 284e290 287 using the Medicare database from 1991 to 2005. Severe and less in a manner similar to criterion A. However, a significantly higher severe adverse outcomes were identified in patients who under- success rate was observed with trabeculectomy with MMC for went primary trabeculectomy, trabeculectomy with scarring, and criterion B (28% Ahmed group vs 44% trabeculectomy group, aqueous shunt with International Classification of Diseases (ICD-9) p ¼ 0.024), criterion C (19% Ahmed group vs 40% trabeculectomy and CPT codes. The rate of severe adverse outcomes was highest group, p < 0.001), and criterion D (9% Ahmed group vs 22% trabe- among patients who had aqueous shunts relative to trabeculec- culectomy group, p < 0.001). A recent publication highlighted the tomy after 1-year follow-up (2.0% aqueous shunt group, 0.6% degree to which surgical outcomes are affected by the criteria used primary trabeculectomy group, 1.3% trabeculectomy with scarring to define success (Rotchford and King, 2010), and this point is group). Before treatment, patients in the aqueous shunt group were clearly illustrated by the Tran study. Twice as many patients in the more likely to have neovascular glaucoma (p < 0.001), uveitic trabeculectomy group required repeat glaucoma surgery (26 glaucoma (p < 0.001), and a prior claim for a severe (p < 0.001) and patients trabeculectomy group vs 13 patients Ahmed group), and a less severe outcome (p < 0.001). Controlling for these covariates more patients in the Ahmed group had penetrating keratoplasty (9 reduced the differences between groups, but they remained patients Ahmed group vs 2 patients trabeculectomy group). statistically and clinically significant. 4.3. Tube Versus Trabeculectomy (TVT) study 4.2. Ahmed glaucoma valve implant vs trabeculectomy The TVT Study is a multicenter randomized clinical trial Wilson et al. (2000) compared the outcomes of the Ahmed comparing the safety and efficacy of aqueous shunt surgery to glaucoma valve implant and trabeculectomy in a randomized trabeculectomy with MMC in patients with previous ocular surgery. clinical trial involving 117 patients. Mean IOP was lower in the A total of 212 patients who had prior cataract and/or glaucoma trabeculectomy group at 1 year (11.4 mm Hg trabeculectomy group filtering surgery were randomly assigned to treatment with a 350- vs 17.2 mm Hg Ahmed group, p ¼ 0.01), and the Ahmed group mm2 Baerveldt glaucoma implant or trabeculectomy with MMC required more medical therapy at last follow-up (11% trabeculec- (Gedde et al., 2005). A higher success rate (IOP 21 mm Hg and tomy group vs 35% Ahmed group, p ¼ 0.01). The cumulative 20% reduction from baseline, IOP > 5 mm Hg, no additional probability of success (IOP < 21 mm Hg and 15% reduction from glaucoma surgery, and no loss of light perception vision) was seen baseline, IOP > 5 mm Hg, no additional glaucoma surgery, and no after aqueous shunt surgery after 3-years of follow-up (84.9% tube loss of light perception vision) was similar between the two group vs 69.3% trabeculectomy group, p ¼ 0.010) (Gedde et al., treatment groups after 1-year follow-up (83.6% trabeculectomy 2009). Both treatment groups had similar mean IOP (13.0 mm Hg group vs 88.1% Ahmed group, p ¼ 0.43). The study, performed in tube group vs 13.3 mm Hg trabeculectomy group, p ¼ 0.78) and Saudi Arabia and Sri Lanka, included patients with all glaucoma mean number of glaucoma medications (1.3 medications tube types and some eyes that had undergone previous ocular surgery. A group vs 1.0 medications trabeculectomy group, p ¼ 0.30) at 3 follow-up study continued recruitment in Sri Lanka to enroll a total years. Postoperative complications were more common after tra- of 123 patients with primary open-angle glaucoma or angle-closure beculectomy with MMC during the first 3 years of the study (39% glaucoma without prior ocular surgery (Wilson et al., 2003). With tube group vs 60% trabeculectomy group, p ¼ 0.004), but no a mean follow-up of 31 months, no significant differences between significant difference in the rate of serious complications associated the trabeculectomy and Ahmed groups were observed with respect with reoperation and/or vision loss of two or more Snellen lines to the mean IOP (13.6 mm Hg trabeculectomy group vs 13.1 mm Hg was observed between treatment groups (22% tube group vs 27% Ahmed group) and mean number of glaucoma medications (0.93 trabeculectomy group, p ¼ 0.58). medications trabeculectomy group vs 1.13 medications Ahmed The TVT Study challenges several current concepts and other group, p ¼ 0.34) at final follow-up. The cumulative probability of publications comparing trabeculectomy and aqueous shunt success was also similar between treatment groups at last follow- surgery. Based upon a systematic review of the medical literature up (68.1% trabeculectomy group vs 69.8% Ahmed group, p ¼ 0.86). on aqueous shunts, a panel of glaucoma specialists concluded that Tran et al. (2009) presented a retrospective case-controlled low IOP levels usually cannot be attained with these devices and study that compared the outcomes of Ahmed glaucoma valve the IOP typically settles in the high teens postoperatively (Minckler implant surgery and trabeculectomy with MMC. The study included et al., 2008). The TVT Study contradicts these opinions, as evi- 166 patients with primary or secondary open-angle glaucoma with denced by a mean IOP of 13.0 mm Hg and IOP of 14 mm Hg or less in at least 3-year follow-up, and the two treatment groups were 62% of patient in the tube group at 3 years (Gedde et al., 2009). matched with respect to age, preoperative surgery, preoperative Studies by Wilson et al. (2000, 2003) and Tran et al. (2009) found IOP, and preoperative medications. Four criteria were used to similar success rates with the Ahmed glaucoma valve implant and compare the probability of success between the two groups: trabeculectomy. In contrast, the TVT Study found that Baerveldt (A) IOP 21 mm Hg and 15% reduction from baseline; implantation had a higher success rate compared with trabecu- (B) IOP 18 mm Hg and 20% reduction from baseline; lectomy with MMC (Gedde et al., 2007a, 2009). The differences in (C) IOP 15 mm Hg and 25% reduction from baseline; and study findings may relate to differences in study populations, (D) IOP 12 mm Hg and 30% reduction from baseline. Additional success and failure criteria, and retention during follow-up. Addi- criteria for success included no loss of light perception vision, no tionally, the Baerveldt glaucoma implant has an equatorial plate additional glaucoma surgery, and no persistent hypotony with a larger surface area than the Ahmed implant, and evidence (IOP 5 mm Hg). A higher probability of success was observed with suggests that shunts with larger surface-area plates are associated trabeculectomy with MMC compared with Ahmed implantation at with greater IOP reduction (Heuer et al., 1992; Minckler et al., 5 years, which became progressively more statistically significant 2008). Stein et al. (2008) noted a higher rate of adverse outcomes as the success criteria were made more stringent. No significant after aqueous shunt surgery than trabeculectomy. The opposite difference was observed between treatment groups for criterion A observation was made in the TVT Study, with a higher rate of (36% Ahmed group vs 48% trabeculectomy group, p ¼ 0.094). This postoperative complications after trabeculectomy with MMC rela- result is consistent with the randomized prospective trials by tive to aqueous surgery (Gedde et al., 2007b, 2009) Stein et al. Wilson et al. (2000, 2003) that reported similar success rates with (2008) and the accompanying Editorial (Javitt, 2008) acknowl- the Ahmed implant and trabeculectomy when success was defined edge the limitations of data derived from Medicare claims, 288 S.J. Gedde et al. / Experimental Eye Research 93 (2011) 284e290 including the possibility of misattributing complications from the those at baseline and those of the control eye at all time points fellow eye. Moreover, aqueous shunts have historically been during the study period. The superotemporal area, the site closest to reserved for patients who have failed standard filtering surgery, the tube, showed the greatest decrease in endothelial density at 24 and there are likely differences in case severity among the Medicare months (22.6% superotemporal vs 15.4% central). While this study beneficiaries in the Stein study who underwent aqueous shunt raises concern about corneal endothelial damage with aqueous surgery and trabeculectomy. In contrast, patients who underwent shunts, it is noteworthy that none of the study eyes experienced these two glaucoma procedures in the TVT Study had similar overt corneal edema or a decline in best-corrected visual acuity. characteristics as a result of the randomization process, and Mendrinos et al. (2009) assessed corneal endothelial cell loss and included lower risk patients than have traditionally had aqueous tube position in 10 patients who had placement of an Ahmed shunt surgery (e.g., only prior clear cornea cataract extraction). implant. Central and peripheral corneal endothelial density was measured with confocal microscopy, and imaging of the anterior 5. Complications chamber tube was performed with anterior segment optical coherence tomography (AC-OCT) at 6 months and 1 year post- Aqueous shunts are associated with similar intraoperative and operatively. Mean endothelial cell density decreased 7.9% in the postoperative complications as occurs after trabeculectomy. Addi- central cornea and 7.5% in the peripheral cornea over the 6-month tionally, there are unique surgical complications that may develop study period. No significant association between central and related to implantation of a foreign material. Motility disturbances, peripheral endothelial cell loss and tube position (intracameral tube tube exposure, and corneal endothelial cell loss are important length and distances between the tube and cornea and tube and adverse events following aqueous shunt surgery, and recent studies iris). The absence of a correlation between tube position and corneal have further clarified the clinical features of these complications endothelial cell loss in this study does not exclude the importance of (Lee et al., 2009; Mendrinos et al., 2009; Rauscher et al., 2009; mechanical factors. Tube position relative to the cornea can change Stewart et al., 2010). with eye rubbing and blinking (especially in eyes with low levels of IOP), and this dynamic relationship would not likely be detected 5.1. Motility disturbances with the AC-OCT measurements made in this study.

The incidence of diplopia has been reported to range from 1.4% 6. Surgical technique to 37% (Ayyala et al., 1998; Britt et al., 1999; Dobler-Dixon et al., 1999; Harbick et al., 2006; Huang et al., 1999; Lloyd et al., 1994; Several modifications in surgical technique for aqueous shunt Roy et al., 2001; Smith et al., 1993; Tsai et al., 2003) and persis- implantation have been recently described. These modifications tent postoperative strabismus has ranged from 2.1% to 77% (Britt have been directed toward improving surgical success (DeCroos et al., 1999; Dobler-Dixon et al., 1999; Krishna et al., 2001; Lloyd et al., 2009; Sahiner et al., 2009), reducing postoperative compli- et al., 1994; Roy et al., 2001; Smith et al., 1993) in case series of cations (Bochmann and Azuara-Blanco, 2009; Camejo and Noecker, aqueous shunts. Rauscher et al. (2009) evaluated preoperative and 2008; Prata et al., 2010), and optimizing efficiency and cost (Kahook postoperative motility disturbances after aqueous shunt surgery in and Noecker, 2006; Ollila et al., 2005; Rossiter-Thornton et al., 2010). the setting of a multicenter randomized clinical trial. Patients enrolled in the TVT Study underwent a formal motility examination 6.1. Implant enclosure in a porous membrane at baseline and at the 1-year follow-up visit, as well as at any visit after 3 months during which the patient reported diplopia. The rate DeCroos et al. (2009) developed an Ahmed glaucoma implant in of new postoperative motility disturbances was 9.9% after Baer- which the plate was enclosed in a porous membrane of polytetra- veldt implantation, and new-onset persistent diplopia was repor- fluoroethylene, termed porous retrofitted implant with modified ted in 5% of patients. enclosure (PRIME-Ahmed). The PRIME-Ahmed was compared with the standard Ahmed implant in a rabbit model. Histologic analysis 6 5.2. Tube exposure weeks after surgical implantation demonstrated a thinner fibrous capsule around the PRIME-Ahmed (46.4 microns PRIME-Ahmed vs Stewart et al. (2010) performed a meta-analysis of 38 prior studies 94.9 control, p < 0.001) with more vascularity near the tis- that included 3255 eyes of 3105 patients who underwent aqueous sueematerial interface. It was suggested that a thinner fibrous shunt placement with a mean follow-up of 26.1 months. The overall capsule with enhanced vascularity may reduce aqueous outflow incidence of tube exposure was 2.0% with an average rate of exposure resistance and improve long-term aqueous shunt performance. of 0.09% per month. No difference in the incidence of exposure was seen among the Ahmed, Baerveldt, and Molteno implants. 6.2. Drug-coated aqueous shunt

5.3. Corneal endothelial cell loss Investigators at Tulane University developed a slow-release antifibrotic drug-coated aqueous shunt (Sahiner et al., 2009). MMC Lee et al. (2009) prospectively evaluated the changes in corneal was incorporated into the polymer poly (2-hydroxyethyl methac- endothelial cells in 41 patients who underwent Ahmed implanta- rylate), and release of the drug occurred on contact with water. This tion. Corneal specular microscopy was performed on the superior, drug delivery device was attached to the Ahmed glaucoma valve superotemporal, superonasal, and central areas with a noncontact and implanted in rabbits. Histologic analysis 3 months post- specular microscope before surgery and 1, 6, 12, 18, and 24 months operatively demonstrated a significant reduction in inflammatory postoperatively. The changes at each time point were compared reaction and fibrosis of the MMC-coated Ahmed implants with those at baseline and with those of a control group, which compared with unmodified Ahmed implants. consisted of 20 contralateral glaucomatous eyes treated with anti- glaucoma medications. The average percentage decrease in corneal 6.3. Tube insertion into the ciliary sulcus endothelial count was 5.8% at 1 month, 11.5% at 6 months, 15.3% at 12 months, 16.6% at 18 months, and 18.6% at 24 months after Tube insertion into the anterior chamber during aqueous shunt surgery. The changes were statistically significant compared with surgery has been associated with progressive corneal endothelial S.J. Gedde et al. / Experimental Eye Research 93 (2011) 284e290 289 cell loss and corneal decompensation, especially in eyes with with trabeculectomy. A 6e7 mm long tunnel is developed from the shallow anterior chambers, compromised corneal endothelial, or equator to the limbal scleral flap with a crescent blade. Once the tube previous penetrating keratoplasty. Pars plana tube insertion has been inserted into the scleral tunnel, an entry incision into the maximizes the distance between the tube and cornea, but requires anterior chamber is made under the limbal scleral flap with a 23- that a pars plana vitrectomy be performed with complete removal gauge needle. A retrospective caseecontrol study was conducted of the vitreous. In patients with posterior chamber intraocular lens comparing this new technique in 92 patients with a traditional implants, insertion of the tube into the ciliary sulcus is a surgical technique in 332 patients. A lower rate of tube erosion was observed option that does not require a concurrent vitrectomy. Prata et al. with the new technique (0% new technique vs 4.5% traditional (2010) reported encouraging midterm clinical outcomes with technique, p ¼ 0.038) with a median follow-up of 22 months. Graft- Baerveldt tube insertion into the ciliary sulcus. Reduction of IOP free tube insertion offers the advantages of reduced cost and between 7 and 18 mm Hg was achieved in 82% of patients with increased efficiency, as well as elimination of the risk of transmitting a mean follow-up of 37.2 months. Mean IOP decreased from disease with use of a patch graft (such as prion diseases). 28.4 12.2 mm Hg to 12.1 5.9 mm Hg, and the mean number of glaucoma medications was reduced from 3.4 1.4 to 2.0 1.4. 7. Conclusions Camejo and Noecker (2008) reported a modified technique for ab interno sulcus placement of aqueous shunt tubes. In preparing the Aqueous shunts are being increasingly utilized in the surgical tube, it is trimmed longer (approximately 4 mm anterior to the management of glaucoma as an alternative to trabeculectomy limbus) with a bevel facing down. A viscoelastic agent is injected to (Joshi et al., 2005; Ramulu et al., 2007). Recent studies have expand the sulcus. A 23-gauge needle is introduced through the demonstrated comparable results with trabeculectomy and cornea 180 from the planned tube insertion site, advanced across the aqueous shunts in similar patient groups (Gedde et al, 2007a, 2009; anterior chamber, and passed under the iris and out the eye through Tran et al., 2009; Wilson et al., 2000, 2003). It is desirable to avoid the sclera. The tube is then inserted through the needle track. the hypotony and bleb-related complications associated with tra- beculectomy. However, aqueous shunts have their own unique set 6.4. Transcameral suture to prevent tube-cornea touch of surgical complications, including motility disturbances (Ayyala et al., 1998; Britt et al., 1999; Dobler-Dixon et al., 1999; Harbick Bochmann and Azuara-Blanco (2009) described a new tech- et al., 2006; Huang et al., 1999; Krishna et al., 2001; Lloyd et al., nique to manage tube malposition after aqueous shunt implanta- 1994; Rauscher et al., 2009; Roy et al., 2001; Smith et al., 1993; e tion. A 10 0 polypropylene suture double-armed with 3-inch long Tsai et al., 2003), tube exposure (Stewart et al., 2010), and corneal straight needles was placed transcamerally from limbus to limbus endothelial loss (Lee et al., 2009; Mendrinos et al., 2009). Modifi- running over the tube in a patient with tube-cornea touch. The tube cations in surgical technique have been described in an attempt to was depressed into an optimal position away from the corneal improve surgical success (DeCroos et al., 2009; Sahiner et al., 2009), endothelium for 20 months follow-up. reduce postoperative complications (Bochmann and Azuara- Blanco, 2009; Camejo and Noecker, 2008; Prata et al., 2010), and 6.5. Fibrin glue-assisted aqueous shunt surgery optimize efficiency and cost (Kahook and Noecker, 2006; Ollila et al., 2005; Rossiter-Thornton et al., 2010). New clinical trials fi Kahook and Noecker (2006) evaluated the use of brin glue as data have identified differences in the safety and efficacy of the a suture substitute for portions of aqueous shunt implantation in aqueous shunts in popular use (Budenz et al., 2011; Nassiri et al., e a retrospective case control study. Fourteen consecutive patients 2010). This information should prove valuable to glaucoma fi who underwent aqueous shunt placement with brin glue were surgeons who wish to select the safest and most effective IOP compared with 28 consecutive patients who had aqueous shunt lowering procedure for each individual patient. surgery with traditional suture material. A 250-mm2 Baerveldt glaucoma implant and Tutoplast (Innovative Ophthalmic Products, Costa Mesa, CA) were used in all patients. Fibrin glue was applied to Acknowledgments close the conjunctiva, secure the patch graft, and secure the tube to sclera with the new technique. No significant differences were Drs. Gedde, Heuer, and Parrish are the study chairmen and Dr. observed in IOP, glaucoma medical therapy, and postoperative Budenz is a coinvestigator for the Tube Versus Trabeculectomy fi complications between treatment groups. Conjunctival inflamma- Study, which was supported by research grants from P zer, Abbott tion was more pronounced in the suture group on a 0e4 scale (2.0 Medical Optics, the National Eye Institute, and Research to Prevent in the fibrin glue group vs 2.83 in the suture group, p ¼ 0.002), and Blindness. Dr. Budenz is a study chairman, Dr. Gedde is a coinvesti- the surgical time was less for the fibrin glue group (15.0 min in gator, and Dr. Heuer is a member of the Safety and Data Monitoring fibrin glue group vs 25.9 min suture group, p < 0.001). Committee for the Ahmed Baerveldt Comparison Study, which was supported by grants from New World Medical, the National Eye 6.6. Graft-free tube insertion Institute, and Research to Prevent Blindness. Dr. Parrish is a member of the Scientific Advisory Board of Glaukos Corporation and Innovia. Different techniques have been described for inserting an Publication of this manuscript was supported in part by unre- aqueous shunt tube through a scleral tunnel without a donor patch stricted grants from Research to Prevent Blindness, Inc, New York, graft (Ollila et al., 2005; Rossiter-Thornton et al., 2010). Rossiter- New York to the Department of Ophthalmology, University of Miami, Thornton et al. (2010) reported the use of a bent vitreoretinal Miller School of Medicine, Miami, Florida and the Department of blade introduced through a partial thickness scleral incision Ophthalmology, Medical College of Wisconsin, Milwaukee, Wisconsin. 5e6 mm from the limbus and guided through the superficial scleral into the anterior chamber. A consecutive case series of 34 patients References with this technique for tube insertion found that only 1 patient developed tube erosion during 10 years of follow-up. Ollila et al. Ayyala, R.S., Zurakowski, D., Smith, J.A., et al., 1998. A clinical study of the Ahmed glaucoma valve implant in advanced glaucoma. Ophthalmology 105, 1968e1976. (2005) described another technique for graft-free tube insertion. A Ayyala, R.S., Harman, L.E., Michelini-Norris, M., et al., 1999. Comparison of different 6 4 mm partial thickness scleral flap is dissected at the limbus, as biomaterials for glaucoma drainage devices. Arch. Ophthalmol. 117, 233e236. 290 S.J. Gedde et al. / Experimental Eye Research 93 (2011) 284e290

Ayyala, R.S., Harman, L.E., Flores, A., et al., 2000. Comparison of different bioma- Lloyd, M.A., Baerveldt, G., Fellenbaum, P.S., et al., 1994. Intermediate-term results of terials for glaucoma drainage devices. Part 2. Arch. Ophthalmol. 118, 1081e1084. a randomized clinical trial of the 350- versus the 500-mm2 Baerveldt implant. Ayyala, R.S., Zurakowski, D., Monshizadeh, R., et al., 2002. Comparison of double- Ophthalmology 101, 1456e1463. plate Molteno and Ahmed glaucoma valve in patients with advanced uncon- Mendrinos, E., Dosso, A., Sommerhalder, J., Shaarawy, T., 2009. Coupling of HRT II trolled glaucoma. Ophthalmic Surg. Lasers 33, 94e101. and AS-OCT to evaluate corneal endothelial cell loss and in vivo visualization of Barton, K., Gedde, S.J., Budenz, D.L., et al., 2011. The Ahmed Baerveldt Comparison the Ahmed glaucoma valve implant. Eye 23, 1836e1844. Study: methodology, baseline patient characteristics, and intraoperative Minckler, D.S., Shammas, A., Wilcox, M., Ogden, T.E., 1987. Experimental studies of complications. Ophthalmology 118, 435e442. aqueous filtration using the Molteno implant. Trans. Am. Ophthalmol. Soc. 85, Bochmann, F., Azuara-Blanco, A., 2009. Transcameral suture to prevent tube-corneal 368e392. touch after glaucoma drainage device implantation: a new surgical technique. Minckler, D.S., Francis, B.A., Hodapp, E.A., et al., 2008. Aqueous shunts in glaucoma: J. Glaucoma 18, 576e577. a report by the American Academy of Ophthalmology. Ophthalmology 115, Britt, M.T., LaBree, L.D., Lloyd, M.A., et al., 1999. Randomized clinical trial of the 350- 1089e1098. mm2 versus the 500-mm2 Baerveldt implant: longer term results. Is bigger Molteno, A.C.B., Straughan, J.L., Ancker, E., 1976. Control of bleb fibrosis after glau- better? Ophthalmology 106, 2312e2318. coma surgery by anti-inflammatory agents. S. Afr. Med. J. 50, 881e885. Brown, R.D., Cairns, J.E., 1983. Experience with the Molteno long tube implant. Molteno, A.C.B., 1980. Mechanisms of intraocular inflammation. Trans. Ophthalmol. Trans. Ophthalmol. Soc. U.K 103, 297e312. Soc. N.Z 32, 69e72. Budenz, D.L., Barton, K., Feuer, W.J., et al., 2011. Treatment outcomes in the Ahmed Molteno, A.C., 1987. Difficult glaucomas [letter]. Eye 1, 632e633. Baerveldt Comparison Study after one year of follow-up. Ophthalmology 118, Nassiri, N., Kamali, G., Rahnavardi, M., et al., 2010. Ahmed glaucoma valve and 443e452. single-plate Molteno implants in treatment of refractory glaucoma: a compar- Cairns, J.E.,1983. The Molteno long-tube implant. Trans. Ophthalmol. Soc. U.K 103, 39e41. ative study. Am. J. Ophthalmol. 149, 893e902. Camejo, L., Noecker, R., 2008. Ab interno sulcus placement of glaucoma tube Ollila, M., Falk, Q., Airaksinen, P.J., 2005. Placing the Molteno implant in a long implants. Ophthalmic Surg. Lasers Imaging 39, 434e435. scleral tunnel to prevent postoperative tube exposure. Acta Ophthalmol. Scand. Cantor, L., Burgoyne, J., Sanders, S., et al., 1998. The effect of mitomycin C on Molteno 83, 302e305. implant surgery: a 1-year randomized, masked, prospective study. J. Glaucoma Prata, T.S., Mehta, A., De Moraes, C.G., et al., 2010. Baerveldt glaucoma implant in the 7, 240e246. ciliary sulcus: midterm follow-up. J. Glaucoma 19, 15e18. Chen, P.P., Yamamoto, T., Sawada, A., et al., 1997. Use of antifibrosis agents and Ramulu, P.Y., Corcoran, K.J., Corcoran, S.L., Robin, A.L., 2007. Utilization of various glaucoma drainage devices in the American and Japanese Glaucoma Societies. glaucoma surgeries and procedures in Medicare beneficiaries from 1995 to J. Glaucoma 6, 192e196. 2004. Ophthalmology 114, 2265e2270. Choritz, L., Koynov, K., Renieri, G., et al., 2010. Surface topographies of glaucoma Rauscher, F.M., Gedde, S.J., Schiffman, J.C., et al., 2009. Motility disturbances in the drainage devices and their influence on human Tenon fibroblast adhesion. Tube Versus Trabeculectomy Study during the first year of follow-up. Am. J. Invest. Ophthalmol. Vis. Sci. 51, 4047e4053. Ophthalmol. 147, 458e466. Corcoran, K.J., 2009. The shifting tides of glaucoma surgery. Rev. Ophthalmol. 16 Rossiter-Thornton, L., Azar, D., Leong, J., et al., 2010. Graft-free Molteno tube (9) accessed online at: http://www.revophth.com/index.asp?page¼1_14458. insertion: 10-year outcomes. Br. J. Ophthalmol. 94, 665e666. htmon (18.09.10.). Rotchford, A.P., King, A.J., 2010. Moving the goal posts. Definitions of success after Costa, V.P., Azuara-Blanco, A., Netland, P.A., et al., 2004. Efficacy and safety of glaucoma surgery and their effect on reported outcome. Ophthalmology 117, adjunctive mitomycin C during Ahmed glaucoma valve implantation: 18e23. a prospective randomized clinical trial. Ophthalmology 111, 1071e1076. Roy, S., Ravinet, E., Mermoud, A., 2001. Baerveldt implant in refractory glaucoma: DeBry, P.W., Perkins, T.W., Heatley, G., et al., 2002. Incidence of late-onset bleb- long-term results and factors influencing outcome. Int. Ophthalmol. 24, 93e100. related complications following trabeculectomy with mitomycin. Arch. Oph- Sahiner, N., Kravitz, D.J., Qadir, R., et al., 2009. Creation of a drug-coated glaucoma thalmol. 120, 297e300. drainage device using polymer technology: in vitro and in vivo studies. Arch. DeCroos, F.C., Ahmad, S., Kondo, Y., et al., 2009. Expanded polytetrafluoroethylene Ophthalmol. 127, 448e453. membrane alters tissue response to implanted Ahmed glaucoma valve. Curr. Schwartz, K.S., Lee, R.K., Gedde, S.J., 2006. Glaucoma drainage implants: a critical Eye Res. 34, 562e567. comparison of types. Curr. Opin. Ophthalmol. 17, 181e189. Dobler-Dixon, A.A., Cantor, L.B., Sondhi, N., et al., 1999. Prospective evaluation of Smith, S.L., Starita, R.J., Fellman, R.L., Lynn, J.R., 1993. Early clinical experience with extraocular motility following double-plate Molteno implantation. Arch. Oph- the Baerveldt 350-mm2 glaucoma implant and associated extraocular muscle thalmol. 117, 1155e1160. imbalance. Ophthalmology 100, 914e918. Gedde, S.J., Schiffman, J.C., Feuer, W.J., et al., 2005. The Tube Versus Trabeculectomy Smith, M.F., Doyle, J.W., Sherwood, M.B., 1995. Comparison of the Baerveldt glau- Study: design and baseline characteristics of study patients. Am. J. Ophthalmol. coma implant with the double-plate Molteno drainage implant. Arch. Oph- 140, 275e287. thalmol. 113, 444e447. Gedde, S.J., Schiffman, J.C., Feuer, W.J., et al., 2007a. Treatment outcomes in the Tube Stein, J.D., Ruiz, D., Belsky, D., et al., 2008. Longitudinal rates of postoperative Versus Trabeculectomy Study after one year of follow-up. Am. J. Ophthalmol. adverse outcomes after glaucoma surgery among medicare beneficiaries 1994 143, 9e22. to 2005. Ophthalmology 115, 1109e1116. Gedde, S.J., Herndon, L.W., Brandt, J.D., et al., 2007b. Surgical complications in the Stewart, W.C., Kristoffersen, C.J., Demos, C.M., et al., 2010. Incidence of conjunctival Tube Versus Trabeculectomy Study during the first year of follow-up. Am. J. exposure following drainage device implantation in patients with glaucoma. Ophthalmol. 143, 23e31. Eur. J. Ophthalmol. 20, 124e130. Gedde, S.J., Schiffman, J.C., Feuer, W.J., et al., 2009. Three-year follow-up of the Tube Syed, H.M., Law, S.K., Nam, S.H., et al., 2004. Baerveldt-350 implant versus Ahmed Versus Trabeculectomy Study. Am. J. Ophthalmol. 148, 670e684. valve for refractory glaucoma: a case-controlled comparison. J. Glaucoma 13, Harbick, K.H., Sidoti, P.A., Budenz, D.L., et al., 2006. Outcomes of inferonasal Baer- 38e45. veldt glaucoma drainage implant surgery. J. Glaucoma 15, 7e12. Tran, D.H., Souza, C., Ang, M.J., et al., 2009. Comparison of long-term surgical Heuer, D.K., Lloyd, M.A., Abrams, D.A., et al., 1992. Which is better? One or two? A success of Ahmed valve implant versus trabeculectomy in open-angle glau- randomized clinical trial of single-plate versus double-plate Molteno implanta- coma. Br. J. Ophthalmol. 93, 1504e1509. tion for glaucomas in aphakia and pseudophakia. Ophthalmology 99, 1512e1519. Tsai, J.C., Johnson, C.C., Dietrich, M.S., 2003. The Ahmed shunt versus the Baerveldt Huang, M.C., Netland, P.A., Coleman, A.L., et al., 1999. Intermediate-term clinical expe- shunt for refractory glaucoma: a single-surgeon comparison of outcome. rience with the Ahmed Glaucoma Valve implant. Am. J. Ophthalmol. 127, 27e33. Ophthalmology 110, 1814e1821. Jamil, A.L., Mills, R.P., 2007. Glaucoma tube or trabeculectomy? That is the question. Tsai, J.C., Johnson, C.C., Kammer, J.A., Dietrich, M.S., 2006. The Ahmed shunt versus Am. J. Ophthalmol. 143, 141e142. the Baerveldt shunt for refractory glaucoma. II: longer-term outcomes from Javitt, J.C., 2008. Trab vs. Tube, a 15-year retrospective: what have we learned in 20 a single surgeon. Ophthalmology 113, 913e917. years? Ophthalmology 15, 1107e1108. Wang, J.C., See, J.L., Chew, P.T., 2004. Experience with the use of Baerveldt and Joshi, A.B., Parrish, R.K., Feuer, W.F., 2005. 2002 survey of the American Glaucoma Ahmed glaucoma drainage implants in an Asian population. Ophthalmology Society. Practice preferences for glaucoma surgery and antifibrotic use. 111, 1383e1388. J. Glaucoma 14, 172e174. Wilcox, M.J., Minckler, D.S., Ogden, T.E., 1994. Pathophysiology of artificial aqueous Kahook, M.Y., Noecker, R.J., 2006. Fibrin glue-assisted glaucoma drainage device drainage in primate eyes with Molteno implants. J. Glaucoma 3, 140e151. surgery. Br. J. Ophthalmol. 90, 1486e1489. Wilson, M.R., Mendis, U., Smith, S.D., Paliwal, A., 2000. Ahmed glaucoma valve Krishna, R., Godfrey, D.G., Budenz, D.L., et al., 2001. Intermediate-term outcomes of implant vs. trabeculectomy in the surgical treatment of glaucoma: a random- 350-mm2 Baerveldt glaucoma implants. Ophthalmology 108, 621e626. ized clinical trial. Am. J. Ophthalmol. 130, 267e273. Lee, E.-K., Yun, Y.-J., Lee, J.-E., et al., 2009. Changes in corneal endothelial cells after Wilson, M.R., Mendis, U., Paliwal, A., Haynatzka, V., 2003. Long-term follow-up of Ahmed glaucoma valve implantation: 2-year follow-up. Am. J. Ophthalmol. 148, primary glaucoma surgery with Ahmed glaucoma valve implant versus trabe- 361e367. culectomy. Am. J. Ophthalmol. 136, 464e470.