Ординатура / Офтальмология / Английские материалы / Mechanisms of the Glaucomas_Shields, Tombran-Tink, Barnstable_2008

glaucomas (1–5). In addition to the advantages of shunts in selected cases, the approximately 5% incidence of late blebitis and endophthalmitis after trabeculectomy with antifibrotic agents has provided added incentive for continued development of artificial drainage devices that function without a conjunctival bleb (6). Historically, the availability of tubes of appropriate diameter and plastic materials with low tissue reactivity were major milestones in the evolution of aqueous shunts. The first successful version of the Molteno implant, introduced in 1969, utilized a silicone tube attached to a round polymethmethacrylate (PMMA)-ridged plate placed equatorially (7,8). The primary tube remains the same in all of the Molteno devices. The double-plate Molteno, now polypropylene with one plate in each of two quadrants usually straddling the superior rectus muscle, became the “gold standard” device to which all subsequent devices have been compared (see Fig. 1) (9). The connecting tube between the plates is identical, except for being shorter, to the primary tube that is placed in the anterior chamber or vitreous cavity. A key feature of the Molteno implants that minimizes fibrous occlusion of the posterior tube orifices is routing the primary and interconnecting tubes through the elevated rim on the explant plates. This feature, copied in the Ahmed and Baerveldt devices (see Table 1), and in the Susanna device (3), ensures that a relatively large separation of the tube opening from the overlying capsule will occur when aqueous flow begins and elevates the capsule off the plate surface to form a filtering bleb. Clinical studies of a newly designed Molteno device (Molteno-3®) have not yet been published. This new design has a lower profile, includes an accessory reservoir designed to obviate temporary ligation, and is implantable in one-quadrant (see Fig. 1).

A remarkable variety of materials have been employed in attempts to achieve lasting translimbal aqueous drainage (anterior shunts) or drainage to equatorial blebs created around equatorially located explant plates (posterior shunts) (1–6).

Besides lumened anterior and posterior shunts, a variety of non-lumened devices (setons) have also been proposed, so far without trial-proven success in humans, to promote aqueous drainage into the supra-choroidal space (1,2).

There is minimal supporting peer-reviewed literature, either experimental or clinical, for the efficacy of several previously described posterior shunts [Schocket (ACTSEB) (10), modified Schocket (11), Joseph (12), OptiMed (13), White shunt-pump (14), Drake-Hoskins (15), e-PTFE (16), Krupin disc (17)]. There is also sparse experimental or clinical peer-reviewed evidence supporting the new ExPRESS™ anterior translimbal shunt (18–21).

The double-plate Molteno implant arguably remains the “gold standard” commercially available aqueous shunt to which others with substantial literature (Ahmed, Baerveldt, Krupin) have been compared. The Molteno has the longest published followup data (1–5,22–25). Exact yearly utilization numbers for aqueous shunts of all types in North America or worldwide are not available but surely number in the tens of thousands. During the last few years, the Ahmed and Baerveldt implants may have superseded the double-plate Molteno in numbers being utilized, principally because they require only one-quadrant surgery. Only time will tell whether the newer Molteno- 3® will again triumph.

Fig. 1. Illustrations of the Ahmed, Baerveldt, and Molteno shunts in position on the eye. (A) Ahmed single-plate; (B) Baerveldt 250/350 with the wings beneath the adjacent muscles; (C) Molteno single-plate; (D) Molteno single-plate with ridge; (E) double-plate Molteno with connecting tube beneath a rectus muscle; (F) Molteno-3. Patch grafts over the anterior portions of the drainage tubes are indicated (donor-preserved pericardium or sclera are most commonly used).

Aqueous shunts have continued to demonstrate about the same efficacy for intraocular pressure (IOP) control as does trabeculectomy with antifibrotics [5- fluorouracil (5-FU), MMC] in complicated glaucomas (9,26,27). One-year success rates have been reported as 70–80+%, dropping to 40–50% by 5 years (3–5). However, very favorable long-term case series by Molteno et al. have demonstrated IOP control for several decades after installation (23–25).

Web product descriptions:

aAVG: New World Medical: http://www.ahmedvalve.com/products/index.html bhttp://www.baerveldt.com/implants/bg-102-350.asp

cMolteno: http://www.molteno.com

PATHOPHYSIOLOGY

Features common to all currently commercially available devices (Ahmed, Baerveldt, Molteno) include explant construction from materials to which fibroblasts cannot tightly adhere (polypropylene, PMMA, silicone rubber) and identical silicone rubber tubes connecting the explant plate to the anterior chamber or vitreous cavity (13). Differences between devices include variations in design, explant surface area and shape, the presence or absence of valves or flow restrictors, and details of surgical installation (3–5,13,28). All shunts currently in wide use with published encouraging shortto medium-term success rates share a ridged equatorial explant plate and a common pathophysiology, which includes stimulating a fibrous capsule around the plate that functions as a biological flow restrictor (13,28). Incorporation of mechanical valves or flow restrictors in aqueous shunts, in theory, should reduce the risk of immediate postoperative hypotony common to non-valved devices when placed without temporary tube ligation. However, immediate postoperative hypotony may occur for reasons besides the lack of a flow restrictor or valve including leak around the tube as it crosses the sclera or decreased aqueous production. Once the capsule around the explant plate is formed and matured, upstream valves or flow restrictors are assumed to be superfluous, a concept recently challenged in bench-modeling experiments (29). A mechanical valve or flow restrictor that limits aqueous outflow would logically increase the risk of obstruction should intraocular bleeding or inflammation and fibrin deposition occur.

Experimental studies including perfusion experiments in monkeys and rabbits have demonstrated that the capsule around the explant provides the primary resistance to aqueous outflow through aqueous shunts (13,30). The standard tube (about 300 μm internal diameter) contributes no measurable resistance to outflow with physiologic perfusion flow rates of 1–2 μl/min. Explant size (mm2) has been shown clinically (9) and experimentally to correlate with drainage capacity of the devices. Capsule thickness also correlates in experimental studies (rabbits) with capsule hydraulic conductivity (flow in μl/min/mmHg/mm2) (31).

Aqueous moves through the capsule into surrounding tissues by simple passive diffusion, demonstrated by horseradish peroxidase (HRP) and latex particles used as markers of flow (30,31). In monkey experiments, latex particles with diameters as large as 1.0 μm diffused rapidly although mature capsules around Molteno implant into surrounding orbital tissues (30). In human specimens in which blood had reached the filtering bleb around the explant, fragments of red cells (±4–5 μm) were found permeating the capsule wall, indicating that relatively large particles can move through the structure. Transmission electron microscopy of capsule walls after perfusion with HRP demonstrated percolation of the marker between bundles of collagen (30).

Conjunctival microcysts rarely develop over the filtering bleb around explants, implying little or no transconjunctival flow, unlike typical peri-limbal trabeculectomy blebs. Generally, the IOP will settle at higher levels (±18 mmHg) than after standard trabeculectomy (14–16 mmHg) or after trabeculectomy with 5-FU or MMC (8–10 mmHg). Flow in non-valved shunts is bi-directional correlating with the ocular pulse, clearly visible by slit lamp when fibrin strands attached to the tube tip move in both directions. Clearance of aqueous from the capsule or from periocular tissues is presumed to occur primarily through venous capillaries (30). Lymphatics, said not to be present in orbital tissues by some authorities, probably play a minor if any role in aqueous clearance out of shunt blebs.

If a very low IOP goal range is desirable, a shunt may be a poor choice for IOP control. However, especially large surface area shunts may rarely be associated with chronic hypotony, even after discontinuing all topical medications. The Baerveldt devices (250 or 350 mm2) can easily be surgically revised without removal to decrease the explant surface area, by cutting one or both wings to a smaller size.

Case reports indicate that intraocular silicone oil can reach the sub-conjunctival tissues and incite inflammation after Ahmed implantation and even after previously placed Ahmed or Baerveldt devices would already be well-encapsulated (32–34).

Recent experimental evidence suggests that the diameter (geometry) of the filtering bleb may play an important role in how encapsulation evolves and how effectively the capsule can drain aqueous (35). Thinner capsules would logically facilitate higher outflow.

EXPLANT ENCAPSULATION AND CORRELATIONS WITH IOP AND FUNCTION

Histological examination of various devices in humans and in animals postoperatively has demonstrated similar capsules around all types of shunt explants. In rabbits and monkeys, inflammatory cell infiltrates are prominent especially in the outer layers

of the capsules both acutely and chronically. In humans, inflammation subsides over time such that almost no inflammatory reaction is found in long-mature functioning capsules (Molteno, personal communication). In rare human cases, in which surgical exploration occurred within weeks of implantation, a distinct capsule has always been present, although noticeably less thick than after several additional weeks or, in failed cases, re-operated after long intervals. When examined using scanning electron microscopy, the internal surface of explant capsules in monkey and man are typically an open collagenous mesh with a scattered population of fibroblasts along the inner surface (30). Within the wall of the capsule a dense layer of organized fibrous tissue is always present, transitioning into a fibrovascular layer near the external surface. Detailed histological and immunohistochemical studies have demonstrated large numbers of proliferating (activated) cells and others undergoing apoptosis, layered so as to imply a continual remodeling of the capsule characterized by apoptosis in deeper layers and renewal from the outside toward the inside (36,37).

With Ahmed or Krupin implants, IOP typically drops immediately postoperatively to relatively low levels (4–6 mmHg) with sub-conjunctival bleb formation (38,39). With ligatured Molteno or Baerveldt implants, the IOP immediately postoperatively will vary depending on surgical technique (28,40,41). Importantly, postoperative hypotony may occur with any of these devices, presumably because of leak around the trans-limbal tubes, valve resistance and tissue variations, or failure of the ligature to completely close the tube with non-valved devices. Perforating the non-valved tube anterior to the ligature may produce variable flow through the tube wall but can also result in hypotony (42–44). Viscoelastic, if used intraoperatively, may moderate IOP drop with either valved or non-valved ligatured tubes, but can also aggravate postoperative IOP rise especially with ligatured tubes. A paracentesis is always advisable intraoperatively to enable easy venting of the anterior chamber if necessary postoperatively and complete removal of viscoelastic is desirable if employed during installation of ligatured tubes.

Characteristic clinical stages will follow installation of aqueous shunt devices beginning after opening the shunt device to aqueous flow or immediate functioning with a valved device. An immediate “hypotensive phase” (usually hours to days) may ensue during which the risks of choroidal effusion or choroidal hemorrhage increase (5,40). Subsequently, a “hypertensive phase” (weeks to months) will then follow typically with eventual resolution to upper teen IOPs with disappearance of inflammatory signs (5,40). Topical medication can be started or increased as necessary when IOP rises and tapered in favorable cases after a “controlled phase” ensues when IOP stabilizes lasting months to years. The presumption has been that the hypertensive phase is due to inflammation provoked by aqueous flow into the capsule and surrounding tissues. Logical management of the hypertensive phase includes topical steroids or other anti-inflammatory medications and resumption of topical or systemic anti-glaucoma medications as required. High IOP is presumed to be a stimulus to excessive fibrosis in the bleb wall. Frequent paracentesis may be necessary during the immediate postoperative phase if IOP rises to unacceptable levels associated with a cystic shunt bleb. Needling the bleb early or late has been recommended by some (J. Freedman, personal communication) but has obvious risks if an easily perforated

silicone explant were implanted. Excision of the capsule after clinical failure may also result in improved IOP control, but the benefits may be transient (42–44).

EXPLANT SURFACE AREA

One prospective, randomized clinical study comparing single-plate Molteno with double-plate Molteno implants in complicated glaucomas supported the importance of surface area to IOP outcome and showed success rates with double-plate Moltenos comparable with those after trabeculectomy with adjunctive 5-FU in similar patients over similar follow-up periods (9,26). Subsequent additional randomized trials comparing two different-sized Baerveldt implants (350 vs. 500 mm2) indicated no clinically significant benefits of 500 mm2 compared with 350 mm2 Baerveldt shunts (45,46). During a subsequent Cochrane Review on aqueous shunts, it was considered appropriate to combine these two latter studies as they represented serial measurements on many of the same patients (47). The validity of the greater efficacy of the doubleplate Molteno versus the single-plate Molteno was less clear utilizing the more rigorous Cochrane methodology, which indicated that the increased efficacy of the larger surface area was statistically significant only in the sub-group of patients considered to be qualified successes who were on adjuvant topical medications (see Table 2) (47).

Some non-randomized clinical series have indicated that larger explant surface area implants (Baerveldt 350 vs. Ahmed single-plate) do provide lower, and longer-term IOP control (48,49). Other studies however have found no difference in IOP outcomes over equivalent time periods comparing single-plate Ahmeds with larger surface area Baerveldts (50).

Experimental data in monkeys and rabbits have supported the importance of surface area to hydraulic conductivity (μl of fluid moving through the capsule per mmHg per minute at physiologic IOP and flow rates) (30,31).

A systematic review of aqueous shunt publications concluded that lower IOPs resulted with larger surface area devices (4). However, retrospective case series continue to report relatively favorable outcomes with small devices, some including long-term follow-up (4,50).

VALVED VERSUS NON-VALVED DEVICES

The attraction of including valve components in aqueous shunts relates to simplifying the surgical installation and avoiding immediate hypotony, expected with non-ligatured or flow-restricted Molteno or Baerveldt devices. Companion in vitro and in vivo experimental studies in rabbits indicated that the valves in the Krupin and Ahmed devices function more as flow restrictors than as conventional on or off valves under physiologic conditions (13). Some reported case series indicate lower immediate hypotony rates with the Ahmed shunt than with non-valved devices (4). Other careful in vitro studies of then available devices have concluded that the Ahmed demonstrates a valvelike function over a wide range of pressures by decreasing or increasing resistance proportional to flow rates (51–53).

Table 2

Randomized Clinical Trials and Aqueous Shunts

ULTRASOUND AND AQUEOUS SHUNTS

B-scan ultrasound is useful to confirm the presence or non-presence of fluid (bleb) over the explant plate, especially when direct visualization of a sub-conjunctival bleb is not possible (54). The ease of clinical visualization of a functioning bleb obviously depends on how posteriorly the explant was placed and on the eye’s mobility and the patient’s orbital anatomy. With a fully functioning shunt, B-scan typically shows flattening of the eye wall between the bleb and the vitreous cavity. Often, ultrasound will reveal a thin layer of fluid over an explant even when IOP is unsatisfactory, most often related to excessive capsule thickness. With Baerveldt or Molteno implants particularly, the explant plate “floats” within an expanded encircling capsule (40).

ANTERIOR CHAMBER VERSUS PARS PLANA TUBE INSERTION

Aqueous shunts may be installed in the anterior chamber regardless of lens status (phakic, pseudophakic, aphakic) or through the pars plana in previously vitrectomized aphakic or pseudophakic eyes or through the posterior chamber in aphakic or pseudophakic eyes in which the iris is pulled forward. Tubes should not be installed in the anterior chamber if corneal or lens touch is likely. It is inadvisable to place tubes in the vitreous cavity unless complete vitrectomy has been done as vitreous may block the tube. Occasionally, after “complete” vitrectomy, residual cortical vitreous may separate and obstruct tubes in the vitreous cavity even years after installation. Clearance of the vitreous plug from the tube may occur after simple massage of the bleb (reversing the pressure gradient) or require aspiration through a vitrectomy instrument.

Whether pars plan tube insertion decreases long-term risk to the cornea or improves penetrating graft survival remains unclear. No randomized trials have been reported. Retrospective studies have indicated graft failure is increased with anterior chamber tube insertion (55,56).

The Hoffman Elbow® is available with the 350 Baerveldt, and a Pars Plana Clip® is available with the Ahmed. Both of these modifications to facilitate pars plana tube insertion have the disadvantage of requiring a large bore needle puncture for insertion (18 gauge) because of the sheath diameter, and both result in higher elevation of the overlying patching material and conjunctiva. In concept, both permit use of pars plana tube insertion with phakic eyes after lens-sparing vitrectomy in that the posterior lens surface is protected from tube contact by the sharp angle of insertion into the mid-vitreous.

SCHOCKET AND MODIFIED SCHOCKET

Besides commercially available shunt devices, a Schocket implant may be assembled from retinal buckle elements and appropriately sized silicone tubing (see Fig. 4) (10). The term “Modified Schocket” has been used to describe utilization of the capsule surrounding a pre-existing buckle element for aqueous drainage (11). Either anterior chamber or pars plana tube insertion is feasible. Assuming a pre-existing encircling buckle of at least a few millimeters width has been in place for even a few weeks, a small segment of silicone tubing may be installed into the surrounding capsule to create an aqueous shunt. Smaller diameter silicone tubing than that included in commercial