Partial thickness corneal tissue as a patch graft material for prevention of glaucoma drainage device exposure

Partial thickness corneal patch grafts covering GDD tubes have favorable long-term outcomes. According to our results, these grafts are associated with a 6.7 % rate of corneal graft melting and a 2.2 % rate of tube exposure. Only one case necessitated additional surgery to cover the exposed tube. A GDD transports aqueous humor from the eye through a tube to an episcleral plate located in one of the quadrants of the globe. The aqueous diffuses through the capsular wall and is absorbed by periocular lymphatics and capillaries [3, 9]. Post-surgical complications unique to GDDs include tube-endothelial contact which may cause significant corneal endothelial damage, fibrous encapsulation leading to filtration failure, continuing low-grade inflammation, and tube or plate migration or extrusion [2].

Conjunctival retraction leading to patch graft exposure is usually considered a minor complication, as long as the tube is well covered by the patch graft [10]. This condition, which does not require surgical intervention, was previously found in 33.5 % of eyes undergoing GDD implantation [10], compared with 8.9 % in the current study. Nevertheless, if the conjunctival retraction is significant and the tube is exposed, this may place the eye at risk for infection, especially if Seidel test is positive. Repairing the exposure in these cases may be challenging, especially if the conjunctiva is scarred.

Direct contact between the conjunctiva and the tube will cause the tube to erode the overlying conjunctiva over time and become exposed [7], putting the eye at risk of infection and endophthalmitis [3]. The most vulnerable site for tube erosion is at its entrance through the sclera, and the role of the patch graft is to prevent tube erosion [3]. Although this approach has been shown to decrease exposure rates, it does not always prevent this complication [7]. The reported rates for tube exposure or device extrusion range between 3 and 30.5 % [7, 10‐13], putting our 2.2 % rate at the bottom of the scale. In the event of tube or plate exposure, the condition of the conjunctiva and patch graft needs to be assessed. The majority of those cases are effectively managed by conjunctival suturing with patch grafting, while poor conjunctiva quality may necessitate the use of a conjunctival auto-graft, amniotic membrane or buccal graft [14, 15]. Re-patching with a different material should be considered in the event that the patch graft has melted [16]. However, removal of the plate or tube may nevertheless be necessary [13] as had been the case in one of our subjects. Causes for tube exposure have not been clearly elucidated in the literature [17], and the reported risk factors are inconsistent. Potential risk might be related to conjunctival status (as affected by aging, inflammation, and device mobility) [18], the number of previous ocular surgeries [18], the number of preoperative hypotensive medications [10], and a previous trabeculectomy [17]. Inferior quadrant implantation bears a higher risk of conjunctival erosion over the plate, possibly because of exposure or a shallower space within the inferior fornix and less conjunctiva and Tenon’s capsule for implant coverage [10, 19, 20]. Another report suggested that tube exposure does not differ between Ahmed, Baerveldt, and Molteno implants [21]. However, since that report included 38 studies with many surgeons using different surgical techniques and different patch grafts, it is difficult to arrive at any firm conclusion from those data. The low rate of complications relevant to the conjunctiva and the partial thickness corneal graft in the current study makes it impossible for us to draw conclusions about the risk factors for graft patch melting and tube exposure. Two out of the 3 cases with melting had the implant placed inferiorly. Also, an immune-mediated process with resultant rapid melting may have contributed to the corneal melting in one of our patients who had underlying uveitis.

Considerations in the choice of patch graft material include biocompatibility, availability, immunologic safety, ease of use, cost and cosmetic appearance [22]. The use of allograft materials, among which the most common are sclera, pericardium, dura mater and cornea, has some disadvantages. Insofar as they are foreign bodies, they have a prima facie potential to erode the overlying conjunctiva [7]. Secondly, they involve considerable cost [7]. The main disadvantage of preserved sclera is its thickness [5], which may be an issue of cosmetic appearance. Sterility and variable quality are other concerns [5]. When an autologous scleral patch graft is used, large areas of sclera may be thinned for harvesting the graft, thus carrying the risk of perforation as well [22]. As for a pericardium graft, it may be a priori too thin and also prone to more melting with longer follow-up [5]. Both pericardium and dura matter are expensive [5]. The use of gamma-irradiated cornea (VisionGraft) for the coverage of a glaucoma tube shunt was recently described as allowing for both decreased risk of disease transmission and improved availability [22]. However, prions are still a risk in these corneas and the high cost is a major obstacle. These drawbacks have led some investigators to seek other solutions, such as placing the tube in a scleral tunnel [8] with or without Tenon advancement and a conjunctival–Tenon flap [7]. These descriptions, however, are all derived from anecdotal reports and the conventional surgical method continues to be the use of a patch graft in GDD surgeries.

The biologic material used as a patch graft in GDD surgeries has at least some role in preventing tube erosion [6], but discussions of these issues are sparse. One report [5] stated that no material was more prone to erosion or melting than another when comparing donor sclera, dura mater and pericardium. Those authors, however, reported a high incidence of donor patch graft thinning (22–26 %) for all 3 of those materials. A corneal patch graft may have some advantages over other covering materials. Its tissue strength and rigidity make it particularly suitable for tectonic support of the ocular wall, and it may be less prone to melting compared to other patch grafts [4, 22, 23]. Moreover, its translucency gives the patient’s eye a better cosmetic appearance [23], which is most relevant in cases where the GDD is placed in the inferior quadrants [22]. Migration, retraction and twisting are possible complications of the tube [24, 25], and they may be difficult to diagnose under an opaque patch graft, such as sclera or pericardium. The translucency of the cornea allows direct visualization of the underlying tube and greater facility in the diagnosis of possible complications [4]. It also facilitates laser suture lysis of the tying suture with non-valved tubes. We believe that the use of a partial thickness corneal graft offers satisfactory tectonic support and has some additional benefits over full thickness corneal grafts. The thin patch we use occupies little of the narrow space of the conjunctiva-sclera, thus minimizing the likelihood of the patch graft to erode the overlying conjunctiva. Indeed, none of our patients had that complication. The thin patch may also reduce the postoperative risk of dellen formation. Finally, the partial thickness corneal graft can be taken from a previous DSEK procedure or from a corneal graft that is unsuitable for optic corneal transplantation. This increases the availability of banked tissue without additional costs.

While a patch graft must be completely covered by the conjunctiva during the surgery, conjunctival laceration or button-hole formation may occur [10]. Simple suturing of the conjunctiva must then be carried out to ensure that the patch graft, tube and plate are entirely covered. This intra-operative complication occurred in one of our cases and was successfully dealt by prompt re-approximation of the conjunctival edges.

Limitations of this study include its retrospective nature, the relative small sample size and the lack of a control group with which to compare our results. Prospective randomized studies comparing partial thickness cornea to other patch graft materials are needed in order to determine which patch is less prone to conjunctival erosion and subsequent tube exposure. Cost considerations, the ideal thickness of the patch graft and its cosmetic appearance should also be taken into consideration. Since our analysis included only patients with follow-up of ≥12 months, complications occurring in patients with shorter follow-up were not reported, and this could have been a confounding factor. Although the mean postoperative follow-up time was more than 2 years, it was only 12 months in four patients. It is not clear whether a longer follow-up would reveal more cases of conjunctival dehiscence or corneal graft melting, but we consider it most probable that more of them would emerge over time. Some authors argued that these complications usually occur within the first 15-month postoperative period [5], while others held that these complications may occur even after 5 years [21]. Graft melting occurred 33 months after surgery in one of our patients, which suggests that patients who underwent GDD surgery should be examined regularly for years after surgery. The strict exclusion criteria that eliminated any cases of previous GDD surgery and patients with past and current ocular surface disease represent a strength of our study by allowing the evaluation of the partial thickness cornea as a patch graft without confounding factors that may influence conjunctival healing or erosion. Another strength is that all surgeries were done by one surgeon using the same technique, unlike most studies that evaluated the results and complications of GDDs placed by different surgeons using different techniques. Indeed, variations in techniques might affect tube exposure [6], making it difficult to draw any conclusions about the effectiveness of a specific patch graft in preventing the silicone tube exposure. We used a corneal patch that had an 8–9 mm diameter. The optimal size of a patch graft that will cover the tube completely from limbus to the plate has not yet been established, and this issue warrants further study.