Glaucoma in Modified Osteo-odontokeratoprosthesis Eyes: Role of Additional Stage 1A and Ahmed Glaucoma Drainage Device–Technique and Timing GEETHA IYER, BHASKAR SRINIVASAN, SHWETA AGARWAL, ROSHNI SHETTY, SRIPRIYA KRISHNAMOORTHY, SHANTHA BALEKUDARU, AND LINGAM VIJAYA PURPOSE:
To report the technique, timing, and outcomes of the Ahmed glaucoma drainage device in eyes with the modified osteo-odonto-keratoprosthesis (MOOKP) and the role of an additional stage 1A to the Rome-Vienna protocol. DESIGN: Retrospective interventional case series. METHODS: Case records of 22 eyes of 20 patients with high intraocular pressure at various stages of the MOOKP procedure performed in 85 eyes of 82 patients were studied. Stage 1A, which includes total iridodialysis, intracapsular cataract extraction, and anterior vitrectomy, was done in all eyes as the primary stage. RESULTS: Seventeen Ahmed glaucoma drainage devices were implanted in 15 eyes of 14 patients (chemical injury in 9 [10 eyes] and Stevens-Johnson syndrome in 5 patients). Implantation was performed during and after stage 1A in 2 and 7 eyes, respectively, after stage 1BD1C in 1 eye, and after stage 2 in 6 eyes. Eleven of 15 eyes (73.3%) remained stable with adequate control of intraocular pressure over a mean follow-up period of 33.68 months (1–90 months). Complications related to the drainage device were hypotony in 1 eye and vitreous block of the tube in 1 eye. CONCLUSION: It is ideal to place the Ahmed glaucoma drainage device prior to the mucosal graft when the anatomy of the ocular surface is least altered with best outcomes. The technique of placement of the drainage device during the various stages of the MOOKP procedure has been described. The intraocular pressure stabilized in three quarters of the eyes with pre-existing glaucoma. (Am J Ophthalmol 2015;159:482–489. Ó 2015 by Elsevier Inc. All rights reserved.)
T
failure. The most common indications for the procedure include chemical injuries and Stevens-Johnson syndrome (SJS). Glaucoma is one of the most common comorbid factors in these eyes and occurs more often with chemical injuries. Management of glaucoma forms an important concern in eyes undergoing any form of keratoprosthesis surgery.1–3 Detection of and monitoring the progression of glaucomatous damage in these eyes is challenging. However with the newer techniques of optic disc imaging and nerve fiber layer analysis, as well as Humphrey visual field testing, progression can be documented reliably in these eyes. Since available tonometers are not reliable to measure the intraocular pressure (IOP) in eyes with keratoprosthesis, the mainstay of measuring the IOP in these eyes is by digital tonometry (the pressure is measured using finger tension).4,5 The conventional antiglaucoma procedure of trabeculectomy is not an option with MOOKP. Diode transscleral and endocyclophotocoagulation have been described earlier with their risks and shortcomings.6 Drainage devices therefore form the mainstay of glaucoma management in these eyes.1,7 However, details regarding the technique of shunt procedures using these devices at various stages of the MOOKP surgery have not been described earlier in literature. In this paper we have described the technique and timing as well as analyzed the outcomes of the Ahmed glaucoma drainage device for glaucoma in eyes undergoing the MOOKP procedure. The role of an additional primary stage 1A procedure in contributing toward lowering of IOP as well as deciding the need for an Ahmed glaucoma drainage device has also been highlighted.
HE MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS
(MOOKP) procedure is performed to restore sight in end-stage ocular surface disorders where conventional penetrating keratoplasty would have a high risk of
THE APPROVAL OF THE INSTITUTIONAL REVIEW BOARD OF
Accepted for publication Nov 19, 2014. From the Medical Research Foundation, Sankara Nethralaya, Chennai, India. Inquiries to Dr Geetha Iyer, Dr G Sitalakshmi Memorial Clinic for Ocular Surface Disorders, C J Shah Cornea Services, Medical Research Foundation, Sankara Nethralaya, 18, College Road, Chennai - 600 006, India; e-mail: [email protected]
Vision Research Foundation, Chennai, India was obtained for this study, which adhered to the tenets of the Declaration of Helsinki. Case records of patients who underwent Ahmed glaucoma drainage device implantation for glaucoma prior to, during, or following the various stages of the MOOKP procedure were analyzed retrospectively.
482
Ó
2015 BY
PATIENTS AND METHODS
ELSEVIER INC. ALL
RIGHTS RESERVED.
0002-9394/$36.00 http://dx.doi.org/10.1016/j.ajo.2014.11.030
Based on the digital tension, the need for maximum systemic acetazolamide, and the optic disc features noted intraoperatively during stage 1A of the MOOKP procedure, the timing of the Ahmed glaucoma drainage device placement was decided. Preoperative evaluation included a complete ophthalmologic evaluation, including ultrasound biomicroscopy, if possible, to evaluate anterior chamber and angle details. The MOOKP at our center broadly follows the guidelines and stages described in the Rome-Vienna Protocol.1 In addition, we routinely perform stage 1A prior to the other 2 stages described in the protocol. STAGES OF MODIFIED OSTEO-ODONTO-KERATOPROS THESIS PROCEDURE: Stage 1A, the additional primary
stage routinely performed at our center, includes total iridodialysis, intracapsular cataract extraction, and anterior vitrectomy. A tectonic penetrating keratoplasty is performed in eyes with corneal thinning or perforation. Fundus evaluation by means of indirect ophthalmoscopy is performed prior to removal of the crystalline lens to assess the status of the retina as well as the optic nerve head. Stage 1BþC, done 1 month following stage 1A, includes placing the mucous membrane graft over the ocular surface, fashioning the osteo-odonto-alveolar lamina, and placing it in the subcutaneous pouch in the cheek. Stage 2 is performed 3 months later. The lamina is placed in the eye and the mucous membrane is placed back, allowing the cylinder to protrude through a central opening made in the mucosa. The placement of the Ahmed glaucoma drainage device can be performed along with or following any of the above-mentioned stages. DEFINITION OF GLAUCOMA:
Glaucoma was diagnosed based on 1 or more of the following factors: a vertical cup-to-disc ratio of 0.7 or more and/or any notching noted during stage 1A; the preoperative digital tension; and the preoperative need for use of systemic acetazolamide. Post stage 2, glaucoma was defined based on the structural appearance as mentioned above and the visual field defect corresponding to the optic nerve changes. Assessment of cup/disc asymmetry is usually not possible in these bilateral end-stage disorders, owing to nonvisualization of the disc in the other eye, and hence cannot be used as a criterion to define glaucoma. Progression was defined as an increase in the cup-to-disc ratio by 0.2 or more with an increase in the visual field defect, with raised digital tension. In eyes with advanced cupping, only the visual field changes and digital tension provided a clue toward progression.
SURGICAL TECHNIQUE:
Placement of ahmed glaucoma drainage device before or after stage 1A. A 360-degree conjunctival peritomy is performed. The superior and lateral rectus muscles are tagged to expose the superotemporal quadrant. The limbus is usually identified in most cases. The plate of the Ahmed glaucoma drainage device is placed deep in the superotemporal fornix and is
VOL. 159, NO. 3
secured 8 mm from the limbus using 8-0 nylon sutures. The tube length is adjusted to protrude3–4 mm beyond the limbus and cut accordingly in a bevel-up fashion. The anterior chamber is entered at the limbus using a 23 or 24 gauge needle and the tube is inserted into the anterior chamber through the entry site. The tube is secured midway to the sclera using 8-0 nylon sutures. The tube is covered using a scleral patch graft that is secured to the sclera using fibrin glue. The conjunctival edges are opposed to each other (in cases of severe forniceal shortening) or secured at the limbus using 8-0 vicryl sutures. Tarsorrhaphy sutures using 5-0 mersilk are placed laterally and/or medially to prevent retraction of the conjunctiva in eyes with shortened fornices. Placement of ahmed glaucoma drainage device following stage 1B. A full-thickness curvilinear incision is placed on the mucosa extending on either side of the superior rectus and the lateral rectus for 4–5 clock hours. Hemostasis is achieved with constant suction and judicious use of cautery. The buccal mucosa is dissected off the sclera and reflected back to expose the limbus. The superior and lateral rectus muscles are tagged. The superotemporal quadrant is exposed adequately for placement of the Ahmed glaucoma drainage device as described above until placement of the scleral patch graft to cover the tube. The buccal mucosa is reflected back over the scleral graft and the incised edges are sutured using 7-0 vicryl absorbable sutures. Placement of ahmed glaucoma drainage device following stage 2. A full-thickness curvilinear incision is placed on the mucosa extending from medial to the superior rectus to inferior to the lateral rectus for 4–5 clock hours. Hemostasis is achieved with constant suction and cautious use of cautery. The buccal mucosa is dissected off the sclera and reflected back to expose the lamina edge. The superior and lateral rectus muscles are tagged if required. The distance of the lamina edge from the center of the cylinder is measured and the anterior chamber entry is made flush along or just anterior to the edge of the lamina. The entry site is made within 9 mm from the center of the lamina so as to be anterior to the pars plana. The superotemporal quadrant is exposed adequately for placement of the device plate as described above until placement of the scleral patch graft to cover the tube. The presence of the tube in the eye is confirmed by visualizing the tube edge beneath the cylinder if possible. The buccal mucosa is reflected back over the scleral graft and the incised edges are sutured using 7-0 vicryl absorbable sutures (Figure 1). A schematic representation of the Ahmed glaucoma drainage device following the 3 stages is shown in Figure 2. POST–MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS FOLLOW-UP PROTOCOL: Topical lubricants and antibiotic
ointment were continued indefinitely. Systemic acetazolamide
GLAUCOMA IN MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS
483
FIGURE 1. Intraoperative pictures of Ahmed glaucoma drainage device placement following stage 2 of the modified osteo-odontokeratoprosthesis procedure. (Left) Mucosa is incised in the superotemporal quadrant close to insertion of the superior rectus and reflected off the sclera up to the edge of the osteo-odonto-alveolar lamina. The plate of the Ahmed glaucoma valve is pushed into the fornix and secured to the sclera using 8-0 nylon sutures. Distance from the center of the cylinder to entry of the tube into the eye should be less than 8–9 mm. (Middle) The end of the tube is visualized inside the eye beneath the optic. (Right) A donor scleral patch graft is placed over the tube along the length of its course, secured with fibrin glue, and the mucosa is reflected back, the edges secured with 7-0 vicryl sutures.
FIGURE 2. Schematic representation of the Ahmed glaucoma drainage device placement following various stages of the modified osteo-odonto-keratoprosthesis procedure. (Left) After stage 1A, the placement of the Ahmed glaucoma drainage device (AGDD) is similar to the placement in a routine glaucomatous eye. The differences include a scarred cornea and conjunctival scarring with shortened fornices. Lateral tarsorrhaphy (LT) performed at the end of the procedure helps prevent conjunctival retraction. (Middle) Placement of the AGDD after Stage 1BDC requires placement of mucosal incision (indicated by dotted line) in the superotemporal quadrant from the superior rectus muscle to the lateral rectus muscle. The mucosa is reflected to expose the limbus and the rest of the surgery is completed. The mucosa is sutured back using 7-0 vicryl sutures. (Right) Placement of the AGDD after stage 2 requires placement of mucosal incision in the superotemporal quadrant from the superior rectus muscle to the lateral rectus muscle. The mucosa is reflected to expose the lamina and the tube is inserted at or as close to the limbus as limited by the edge of the lamina. The point of entry into the eye should not be posterior to 8 mm from the center of the optical cylinder. The mucosa is sutured back using 70 vicryl sutures. SR, superior rectus muscle; IR, inferior rectus muscle; LR, lateral rectus muscle.
was used based on need for up to 6 months after the surgery. The eye was subjected to a thorough clinical examination, posterior segment evaluation with a 90-diopter lens, Humphrey visual field testing every 6 months, B-scan ultrasound, and spiral computed tomography of the lamina once a year. All patients had a descriptive record of the disc details at every visit. A photographic documentation at each visit was not possible. Humphrey visual fields 24-2 and 10-2 were performed in all eyes once every 6 months. 484
RESULTS DEMOGRAPHICS:
Eighty-five eyes of 82 patients underwent the MOOKP procedure at our center between March 2003 and March 2013. Of the 85 eyes that underwent the MOOKP procedure, 44 belonged to the SJS group and 37 to the chemical injury group, with a mean (range) follow-up of 36.93 months (1–90 months) and 30.42 months (1–80), respectively. Four eyes belonged to
AMERICAN JOURNAL OF OPHTHALMOLOGY
MARCH 2015
other etiologies. Twenty-one eyes of 19 patients had glaucoma as defined. One patient was referred to our center for second shunt placement surgery and was included in the study. Of these 20 patients, 15 eyes of 14 patients underwent the Ahmed glaucoma drainage device procedure (17 devices) at various stages of the MOOKP procedure. These included 10 male and 4 female patients with ages ranging from 23 to 54 years. The primary etiology was chemical injury in 9 patients and Stevens-Johnson syndrome in 5 patients. Among the SJS group, 3 had prior intraocular surgeries for corneal perforation and 1 was the patient operated elsewhere for the MOOKP procedure. The Ahmed glaucoma drainage device was implanted in the eye during stage 1A in 2 eyes, following stage 1A in 7 eyes, after stage 1Bþ1C in 1 eye, and after stage 2 in 6 eyes (twice in 1 eye, accounting for 7 valves). The other eye of the same patient had 2 devices implanted: 1following stage 1A and the other subsequent to stage 2. Clinical features, the timing of Ahmed glaucoma drainage device placement, and their outcomes have been listed in the Table.
nerve cupping during stage 1A of the procedure, the patient was keen to go ahead with the rest of the stages of MOOKP in an attempt at visual rehabilitation. However, the visual acuity did not improve beyond perception of light following the MOOKP procedure owing to the glaucomatous optic atrophy and high IOP, despite maximal systemic acetazolamide. Intraoperative findings revealed a normal anterior segment with no peripheral anterior synechiae. There was no previous record available regarding prior use of topical or systemic steroids indicating the possibility of steroid-induced glaucoma. Details of patients who underwent Ahmed glaucoma drainage devices have been tabulated in the Table. Only 2 eyes developed complications related to the Ahmed glaucoma drainage device. In 1 patient, in whom the valve procedure was done along with stage 1A, persistent hypotony occurred and required tube ligation 2 months following the procedure. This resulted in improvement in the IOP, and the rest of the stages could be performed. A best-corrected visual acuity of 20/50 was achieved. In the second eye, vitreous blocked the tube, necessitating an endoscopic vitrectomy.
OUTCOME:
Eleven of 15 eyes (73.3%) remained stable with adequate control of IOP. These 11 patients (11 eyes) continued to receive 125 mg of systemic acetazolamide 2– 3 times a day with potassium supplements. Glaucoma in 2 eyes was well controlled for 6 and 5 years following the MOOKP procedure. However, 1 eye developed endophthalmitis with subsequent phthisis and the other developed a sterile vitritis with laminar resorption, for which the keratoprosthesis was removed. The patient who had the Ahmed glaucoma drainage device placed prior to stage 2 of the procedure (done elsewhere) required a second device for raised IOP 6 months after the MOOKP and the same was placed at our center in the superonasal quadrant. The end of the earlier tube could not be visualized through the optical cylinder. The tube placed at our center developed a vitreous block for which an endoscopic vitrectomy was performed elsewhere, following which the IOP was well maintained. Retinal detachment occurred 1 year after the MOOKP procedure following yttrium-aluminum-garnet membranotomy for retroprosthetic membrane in the 1 eye. Post retinal detachment surgery, the best-corrected visual acuity was 20/400 and has been noted to be stable since then. IOP was well controlled in the remaining 7 eyes. Among the remaining 4 of the 15 eyes, deterioration was noted in the Humphrey visual field along with an increase in the digital tension over the period of follow-up in both eyes of a patient, and a second Ahmed glaucoma drainage device was implanted in the inferotemporal quadrant 6 years after the MOOKP procedure (Figure 3). The IOP was well controlled following the procedure. One patient developed a retinal detachment 3 months following the procedure. The patient was not keen for any further surgical intervention. The fourth patient was a young, 1-eyed patient of SJS and despite having noted the advanced optic
VOL. 159, NO. 3
DISCUSSION GLAUCOMA CONTINUES TO REMAIN ONE OF THE MOST
important sight-threatening concerns following any keratoprosthesis procedure.1,8 Pre-existing glaucoma could be primary or secondary to the etiology or previous intraocular surgical procedures, respectively. It can also occur de novo after the keratoprosthesis is implanted in the eye. Issues in diagnosis, monitoring, and managing glaucoma specific to eyes undergoing the MOOKP surgery are manifold. They are mainly in the IOP measurements, angle assessments, and lack of benefit of topical medications owing to the questionability of penetration of the same through the thick buccal mucosa. Systemic acetazolamide remains the only means to control the IOP when indicated.1 It may not always be possible to assess the anterior segment status by means of ultrasound biomicroscopy, especially in cases of ankyloblepharon or severe forniceal shortening. Despite these limitations, early diagnosis of and intervention for glaucoma in these eyes is of utmost importance. In the earlier long-term study by Falcinelli and associates, 66 of 181 eyes were noted to have pre-existing glaucoma and 10.4% developed de novo glaucoma after the surgery.4 In their systematic review of surgical outcomes of osteoodontokeratoprosthesis,3 Marchi and associates were noted to have reported a 33% occurrence of de novo glaucoma following MOOKP (85 eyes). The Singapore OOKP Study in 2009 reported no occurrence of glaucoma postoperatively with 5 eyes of 16 patients detected to have pre-existing glaucoma. We also noted no occurrence of glaucoma following the procedure in
GLAUCOMA IN MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS
485
486
TABLE. Details of Patients who Underwent the Ahmed Glaucoma Drainage Device Procedure at Different Stages of the Modified Osteo-odonto-keratoprosthesis Procedure
S. No
Age/Sex
Etiology
Prior Intraocular Procedures/Duration at Which AGV Done
Best BCVA
Final BCVA
Disc Statusa— Progression
Follow-up in Months
Comments
With stage 1A 1 31/M 2 24/F
Chemical Chemical
None Bulbar mucosal graft done elsewhere for ocular surface reconstruction
20/50 20/40
20/50 20/40
0.9 0.6
36 24
Hypotony—tube ligated 2 m later UBM—angle closed
After stage 1A 1 24/M 2 35/M
Chemical Chemical
PK þ KLAL 3 m after 1A 9 m after 1A
20/30 20/50
20/30 20/400
0.9 0.6
72 36
AMERICAN JOURNAL OF OPHTHALMOLOGY
7 34/F After stage 1BþC 1 54/M
Chemical
PK 2 years prior/11 m after 1A
PL
20/30
PL
63
Stable UBM—angle closed Thick RCM plastered to iris as seen during stage 1A RD following YAG Anterior segment formed on slit-lamp evaluation Diode CPC 1 m before 1A Stable UBM—angle closed High myopia—RD 3 m after stage 2— refused intervention Well-formed anterior segment noted during 1A GOA noted at 1A ?endophthalmitis-related phthisis
SJS
PK for corneal perforation 6 m prior to 1BþC/2 m after 1BþC
20/150
20/150
0.8
48
Stable
After stage 2 1 33/M
Chemical
20/30
20/60
Adv
84
20/30
20/40
Adv
72
0.9
60
Progression 6 y after 1st valve/stable 1 y after 2nd Progression 6 y after 1st valve/stable 1 y after 2nd Sterile vitritis with laminar resorption— lamina removed Stable
3
33/M
Chemical
Diode CPC 12 m after 1A
20/40
20/40
0.9
72
4 5
41/M 28/M
Chemical SJS
12 m after 1A Diode CPC – twice; 3 m after 1A
20/30 PL
20/30 PL
Adv cupping
67 36
6
33/F
SJS
1 m after 1A
PL
PL
Adv cupping
1
2
33/M
Chemical
Diode CPC/18 m after stage 26 y after stage 2 Diode CPC/6 y after stage 2
3
29/M
Chemical
3 y after stage 2
20/30
PL
4
40/F
SJS
20/80
20/120
0.7–0.9
74
5
25/M
Chemical
Stage 1A done for leaking cystoid cicatrix 3 y prior/3 y after stage 2 2 m after stage 2
20/150
20/150
0.6
48
6
50/M
SJS
2nd valve placed 6 m after stage 2
20/30
20/30
Adv cupping
24
Stable fibrous band from disc cause for reduced BCVA Stable 1st valve placed prior to stage 2— MOOKP surgery done elsewhere
MARCH 2015
Continued on next page
Adv ¼ advanced; AGV ¼ Ahmed glaucoma valve; BCVA ¼ best-corrected visual acuity; CPC ¼ cyclophotocoagulation; GOA ¼ glaucomatous optic atrophy; KLAL ¼ keratolimbal allograft; m ¼ months; PK ¼ penetrating keratoplasty; PL ¼ perception of light; RCM ¼ retrocorneal membrane; RD ¼ retinal detachment; UBM ¼ ultrasound biomicroscopy; y ¼ years; YAG ¼ yttrium-aluminum-garnet. a Disc status mentioned as cup-to-disc ratio.
Adv cupping 0.8 Disc pallor
60 12 36 48 12 0.8
PL 20/130 3/60 PL 20/400 Chemical Chemical Chemical Chemical Chemical 33/M 44/M 40/M 23/M 30/M 3 4 5 6 7
Chemical 24/M
AGV not done 1 40/M
2
Chemical
Diode CPC PK-twice; once for corneal perforation
20/60 20/130 20/400 20/30 20/400
Adv cupping NPL 20/20
5 0.6 NPL 20/30
24
Uncontrolled IOP on acetazolamide developed endophthalmitis in 5 m— removed lamina NPL owing to absolute glaucoma—same lamina shifted to the other eye Laminar resorption—removed Stable Stable Laminar resorption—removed Stable
Comments Follow-up in Months Disc Statusa— Progression Final BCVA Best BCVA Prior Intraocular Procedures/Duration at Which AGV Done Etiology Age/Sex S. No
TABLE. Details of Patients who Underwent the Ahmed Glaucoma Drainage Device Procedure at Different Stages of the Modified Osteo-odonto-keratoprosthesis Procedure (Continued )
VOL. 159, NO. 3
our series (85 eyes). Of these 85 eyes, pre-existing glaucoma was noted in 45.94% of eyes with chemical injury, in comparison to 9.1% of eyes with SJS. The finding of a pale disc in eyes following chemical injuries with a normal finger tension could be attributable to the direct toxic effect of the chemical by intraocular penetration, more commonly alkali, on the retinal ganglion cells, immediately after the injury.8 Glaucoma in SJS eyes was noted in eyes with prior corneal perforation (3 of 4 eyes, 75%; 1 operated for MOOKP elsewhere), leading to damage to both the anterior segment and occlusion of angle structures. In these eyes with pre-existing iridocorneal adhesions, a total iridodialyis is not possible and the iris has to be mechanically abscissed to the extent possible with residual angle closure. In the remaining 40 out of the 44 SJS eyes, the disc was noted to be healthy, with no de novo occurrence of glaucoma following the procedure. Stage 1A described in our series of patients is not a part of the Rome-Vienna protocol. However, performing this stage prior to harvesting the tooth allows intraoperative visualization of the optic disc and retina. This helps assess the possible visual potential of the eye as well as decide the aggressiveness with which to treat the glaucoma. The removal of the iris from its root, the lens extraction by the intracapsular technique, and the anterior vitrectomy done during this stage may play a role in reducing the intraocular pressure in certain cases. The decision to abandon further stages of the MOOKP procedure can be based on the retinal and optic nerve findings during stage 1A of the procedure rather than visualizing the posterior segment for the first time after the completion of the entire MOOKP procedure. In cases of advanced glaucomatous changes with a preoperative high digital tension despite maximum systemic acetazolamide, the Ahmed glaucoma drainage device procedure can be combined with stage 1A of MOOKP. However, combining these can lead to significant hypotony in the postoperative period, as was noted in 1 of our cases. The decision to implant the Ahmed glaucoma drainage device can therefore be taken subsequent to noting the effect of stage 1A on the IOP. Endo-cyclophotocoagulation is an alternative that can be performed during this stage after the placement of a temporary keratoprosthesis with the aid of a vitreoretinal colleague or through endoscopic visualization. If the IOP continues to remain high after stage 1A, the Ahmed glaucoma drainage device is placed prior to stage 1B. This facilitates ease of placement of the device as well as counters the IOP spike because of the mucosal graft. This is indicated if the patient is already on maximum dose of systemic acetazolamide prior to stage 1B of the procedure as it limits options for addressing the postoperative spike to only intravenous mannitol, which is effective for only a few days. Draping the buccal mucosal graft over the entire ocular surface compromises the uveoscleral outflow of aqueous, thereby leading to an increase in the IOP. In 5
GLAUCOMA IN MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS
487
FIGURE 3. 10-2 Humphrey visual field test showing progression of visual field defect between 2008 and 2013, in the (Left) right eye and (Right) left eye of the patient who underwent the modified osteo-odonto-keratoprosthesis procedure in both eyes. Following the progression noted in the Humphrey visual field, a second tube shunt was placed in the inferotemporal quadrant of both the eyes of the patient.
out of 7 eyes, IOP remained stable over the period of followup, in the group wherein the Ahmed glaucoma drainage device was placed before stage 1B. In such cases where the Ahmed glaucoma drainage device is placed prior to buccal mucosal grafting, care has to be taken during subsequent procedures to avoid inadvertent damage to the bleb or the tube while dissecting the conjunctiva or mucosa and tagging the recti muscles. A specific note of the device quadrant should be made in the operative notes. In this context it is important to cover the tube with a scleral patch graft to avoid erosion and exposure of the tube through the mucosa, to prevent inadvertent cutting of the tube during subsequent surgeries, as well as to prevent erosion of the tube by the lamina placed subsequently over it. While performing the Ahmed glaucoma drainage device procedure after stage 2, the point of entry into the eye is important. It is best to be as anterior as possible in order to avoid inadvertent entry through the retina as well as remain parallel and flush with the cornea. This anterior placement of the tube makes the possibility of vitreous blocking the tube remote. The distance from the center of the optical cylinder for tube entry should be anterior to 8–9 mm. Though an endoscopic endocyclophotocoagulation can be an alternate means to a shunt procedure, earlier reports have not been very encouraging. Kumar and associates have described the use of transconjunctival photocoagulation during stage 2 in 3 eyes with glaucoma. Endoscopic diode cyclophotocoagulation was performed in 2 eyes with progression, of which 1 developed endophthalmitis.5 The authors do not have experience with endoscopic endocyclophotocoagulation. Five out of 6 eyes that underwent Ahmed glaucoma drainage device implantation 488
after stage 2 and 6 of 7 eyes that underwent implantation after stage 1A showed advanced glaucomatous damage. Three of the 6 eyes requiring first or second shunt procedure after stage 2 showed a progression in the glaucomatous damage. Only 1 eye required a second shunt for progression when the first was performed after stage 1A. Placing the tube early in the stages of the MOOKP procedure could play a role in preventing the impact of IOP spikes on the nerve during the subsequent procedures as well as longterm progression of glaucomatous damage. There have been reports of glaucoma progression and appearance of de novo glaucoma following any keratoprosthesis procedure. The mechanism for the same following the Boston type 1 keratoprosthesis has been attributed to iridocorneal adhesions and/or inflammatory debris clogging the trabecular meshwork.9 Following the MOOKP procedure wherein a total iridodialysis and intracapsular cataract extraction is performed routinely, there exists very minimal possibility of iridocorneal adhesions forming owing to either the iris or capsular remnant. The anatomic distortion of the angle structures caused by a large-diameter corneal graft in a Boston type 1 keratoprosthesis does not occur in the MOOKP procedure. A central 3.5- to 4-mm opening based on the optical cylinder dimension is the only opening into the eye. There are no peripheral corneal sutures distorting the angle. These could be possible explanations for absence of de novo glaucoma occurring following the MOOKP procedure in our series. The perioperative spike in the intraocular pressure noted in almost all our cases occurs owing to a compromise caused in the uveoscleral outflow of the aqueous by the mucous membrane graft draped over the ocular surface up to the insertion of the recti muscles. This spike is managed by means of intravenous mannitol
AMERICAN JOURNAL OF OPHTHALMOLOGY
MARCH 2015
for 3–4 days in the immediate postoperative period followed by systemic acetazolamide. Systemic acetazolamide is usually continued for a period of 6 months to 1 year following the MOOKP procedure in eyes with no preexisting glaucoma. The means to monitoring the progression of glaucomatous changes is by measuring the digital tension, serial Humphrey visual field examinations, and optic disc photography. Since each stage of the MOOKP procedure causes a varied response on the IOP, the decision to place the shunt should be based on pre- and postprocedure assessment of IOP. The amount of overfiltration might be more controlled with the Ahmed glaucoma drainage device compared to the Baervaldt tube,5 though the authors have no experience with the Baervaldt tube in their series of patients. Tan and associates have described a Baerveldt drainage device without tube connection being placed in 1 eye for future use if required during stage 1 of the procedure.5 From the authors’ experience, it is better to place the drainage device after stage 1A, rather than concurrently, if the digital tension continues to remain high despite maximum systemic acetazolamide treatment. There is no role of topical antiglaucoma medications at this stage as there is very limited penetration through the buccal mucosa later. Mucosal grafting will invariably cause further spike in the IOP. The authors prefer to avoid intraocular
surgery along with placement of mucous membrane graft during stage 1BþC of the procedure owing to the possibility of an infection from the mucous membrane. It might be desirable to leave the intraocular length of the tube long enough to be visualized beneath the optic cylinder to enable visualization of any tube blockage as well as decrease the possibility of tube retraction and thereby failure to function. The need for a second valve (3 eyes in our series [2 eyes after 6 years following detection of progression on fields, and 1 after 6 months]) could be owing to failure of the drainage device after a few years. The location for the second device would preferably be in the inferotemporal quadrant, followed by the superonasal quadrant. Preplacement of the shunt in all cases undergoing the MOOKP might not be warranted considering the absence of de novo glaucoma in our series of patients. There have been earlier studies regarding issues in monitoring the IOP in eyes undergoing the MOOKP procedure as well as studies suggesting the role of the Ahmed glaucoma drainage device in these eyes. The authors believe that this is the first study reporting the outcome of the Ahmed glaucoma drainage device in MOOKP eyes, describing in detail the technique and timing of the same, along with highlighting the role of an additional stage 1A to the MOOKP procedure.
ALL AUTHORS HAVE COMPLETED AND SUBMITTED THE ICMJE FORM FOR DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST and none were reported. The authors indicate no funding support. Contributions of authors: concept and design of the study (G.I., B.S., V.L.); analysis and interpretation (G.I., B.S., V.L.); writing the article (G.I., B.S.); critical revision of the article (G.I., B.S., V.L.); final approval (G.I., B.S., V.L.); data collection (G.I., B.S., S.A., R.S.); provision of materials, patients, or resources (G.I., B.S., S.A., R.S., S.K., S.B., L.V.); statistical expertise (G.I.); literature search (G.I., B.S.). The authors acknowledge the contribution of Dr Parthopratim Dutta Majumder, Associate Consultant, Medical Research Foundation, towards schematic representation of the placement of the Ahmed Glaucoma Drainage Device following various stages of the modified osteo-odonto-keratoprosthesis procedure in the form of line diagrams in Figure 2.
REFERENCES 1. Hille K, Grabner G, Liu C, et al. Standards for modified osteoodontokeratoprosthesis (OOKP) surgery according to Strampelli and Falcinelli: the Rome-Vienna Protocol. Cornea 2005;24(8):895–908. 2. Liu C, Okera S, Tandon R, Herold J, Hull C, Thorp S. Visual rehabilitation in end-stage inflammatory ocular surface disease with the osteo-odonto-keratoprosthesis: results from the UK. Br J Ophthalmol 2008;92(9):1211–1217. 3. Tan A, Tan DT, Tan XW, Mehta JS. Osteo-odonto keratoprosthesis: systematic review of surgical outcomes and complication rates. Ocul Surf 2012;10(1):15–25. 4. Falcinelli GC, Falsini B, Taloni M, Piccardi M, Falcinelli G. Detection of glaucomatous damage in patients with osteoodontokeratoprosthesis. Br J Ophthalmol 1995;79(2):129–134.
VOL. 159, NO. 3
5. Kumar RS, Tan DT, Por YM, et al. Glaucoma management in patients with osteo-odonto-keratoprosthesis (OOKP): the Singapore OOKP Study. J Glaucoma 2009;18(5):354–360. 6. Lee RM, Al Raqqad N, Gomaa A, Steel DH, Bloom PA, Liu CS. Endoscopic cyclophotocoagulation in osteo-odontokeratoprosthesis (OOKP) eyes. J Glaucoma 2011;20(1):68–69. 7. Hille K, Hille A, Ruprecht KW. Medium term results in keratoprostheses with biocompatible and biological haptic. Graefes Arch Clin Exp Ophthalmol 2006;244(6): 696–704. 8. Cade F, Grosskreutz CL, Tauber A, Dohlman CH. Glaucoma in eyes with severe chemical burn, before and after keratoprosthesis. Cornea 2011;30(12):1322–1327. 9. Crnej A, Paschalis EI, Salvador-Culla B, et al. Glaucoma progression and role of glaucoma surgery in patients with Boston keratoprosthesis. Cornea 2014;33(4):349–354.
GLAUCOMA IN MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS
489
Biosketch Geetha Iyer completed her post graduation in Ophthalmology and Cornea Fellowship from and is currently working as Senior Consultant at the prestigious eye institute Sankara Nethralaya, Chennai, India. She specializes in cornea and ocular surface disorders’ management and has so far done more than 100 keratoprosthesis surgeries (MOOKP, Boston Type 1 and 2 Kpro). Her current ocular research is focused on Stevens Johnson syndrome and chemical injuries.
489.e1
AMERICAN JOURNAL OF OPHTHALMOLOGY
MARCH 2015
Biosketch Bhaskar Srinivasan completed his Cornea Fellowship from and is currently working as Senior Consultant at the prestigious eye institute Sankara Nethralaya, Chennai, India. He is a clinician specializing in cornea and ocular surface disorders including keratoprosthesis and his focused areas of interest include chemical injuries to the eye, lamellar and refractive surgeries.
VOL. 159, NO. 3
GLAUCOMA IN MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS
489.e2
To report the technique, timing, and outcomes of the Ahmed glaucoma drainage device in eyes with the modified osteo-odonto-keratoprosthesis (MOOKP) and the role of an additional stage 1A to the Rome-Vienna protocol. DESIGN: Retrospective interventional case series. METHODS: Case records of 22 eyes of 20 patients with high intraocular pressure at various stages of the MOOKP procedure performed in 85 eyes of 82 patients were studied. Stage 1A, which includes total iridodialysis, intracapsular cataract extraction, and anterior vitrectomy, was done in all eyes as the primary stage. RESULTS: Seventeen Ahmed glaucoma drainage devices were implanted in 15 eyes of 14 patients (chemical injury in 9 [10 eyes] and Stevens-Johnson syndrome in 5 patients). Implantation was performed during and after stage 1A in 2 and 7 eyes, respectively, after stage 1BD1C in 1 eye, and after stage 2 in 6 eyes. Eleven of 15 eyes (73.3%) remained stable with adequate control of intraocular pressure over a mean follow-up period of 33.68 months (1–90 months). Complications related to the drainage device were hypotony in 1 eye and vitreous block of the tube in 1 eye. CONCLUSION: It is ideal to place the Ahmed glaucoma drainage device prior to the mucosal graft when the anatomy of the ocular surface is least altered with best outcomes. The technique of placement of the drainage device during the various stages of the MOOKP procedure has been described. The intraocular pressure stabilized in three quarters of the eyes with pre-existing glaucoma. (Am J Ophthalmol 2015;159:482–489. Ó 2015 by Elsevier Inc. All rights reserved.)
T
failure. The most common indications for the procedure include chemical injuries and Stevens-Johnson syndrome (SJS). Glaucoma is one of the most common comorbid factors in these eyes and occurs more often with chemical injuries. Management of glaucoma forms an important concern in eyes undergoing any form of keratoprosthesis surgery.1–3 Detection of and monitoring the progression of glaucomatous damage in these eyes is challenging. However with the newer techniques of optic disc imaging and nerve fiber layer analysis, as well as Humphrey visual field testing, progression can be documented reliably in these eyes. Since available tonometers are not reliable to measure the intraocular pressure (IOP) in eyes with keratoprosthesis, the mainstay of measuring the IOP in these eyes is by digital tonometry (the pressure is measured using finger tension).4,5 The conventional antiglaucoma procedure of trabeculectomy is not an option with MOOKP. Diode transscleral and endocyclophotocoagulation have been described earlier with their risks and shortcomings.6 Drainage devices therefore form the mainstay of glaucoma management in these eyes.1,7 However, details regarding the technique of shunt procedures using these devices at various stages of the MOOKP surgery have not been described earlier in literature. In this paper we have described the technique and timing as well as analyzed the outcomes of the Ahmed glaucoma drainage device for glaucoma in eyes undergoing the MOOKP procedure. The role of an additional primary stage 1A procedure in contributing toward lowering of IOP as well as deciding the need for an Ahmed glaucoma drainage device has also been highlighted.
HE MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS
(MOOKP) procedure is performed to restore sight in end-stage ocular surface disorders where conventional penetrating keratoplasty would have a high risk of
THE APPROVAL OF THE INSTITUTIONAL REVIEW BOARD OF
Accepted for publication Nov 19, 2014. From the Medical Research Foundation, Sankara Nethralaya, Chennai, India. Inquiries to Dr Geetha Iyer, Dr G Sitalakshmi Memorial Clinic for Ocular Surface Disorders, C J Shah Cornea Services, Medical Research Foundation, Sankara Nethralaya, 18, College Road, Chennai - 600 006, India; e-mail: [email protected]
Vision Research Foundation, Chennai, India was obtained for this study, which adhered to the tenets of the Declaration of Helsinki. Case records of patients who underwent Ahmed glaucoma drainage device implantation for glaucoma prior to, during, or following the various stages of the MOOKP procedure were analyzed retrospectively.
482
Ó
2015 BY
PATIENTS AND METHODS
ELSEVIER INC. ALL
RIGHTS RESERVED.
0002-9394/$36.00 http://dx.doi.org/10.1016/j.ajo.2014.11.030
Based on the digital tension, the need for maximum systemic acetazolamide, and the optic disc features noted intraoperatively during stage 1A of the MOOKP procedure, the timing of the Ahmed glaucoma drainage device placement was decided. Preoperative evaluation included a complete ophthalmologic evaluation, including ultrasound biomicroscopy, if possible, to evaluate anterior chamber and angle details. The MOOKP at our center broadly follows the guidelines and stages described in the Rome-Vienna Protocol.1 In addition, we routinely perform stage 1A prior to the other 2 stages described in the protocol. STAGES OF MODIFIED OSTEO-ODONTO-KERATOPROS THESIS PROCEDURE: Stage 1A, the additional primary
stage routinely performed at our center, includes total iridodialysis, intracapsular cataract extraction, and anterior vitrectomy. A tectonic penetrating keratoplasty is performed in eyes with corneal thinning or perforation. Fundus evaluation by means of indirect ophthalmoscopy is performed prior to removal of the crystalline lens to assess the status of the retina as well as the optic nerve head. Stage 1BþC, done 1 month following stage 1A, includes placing the mucous membrane graft over the ocular surface, fashioning the osteo-odonto-alveolar lamina, and placing it in the subcutaneous pouch in the cheek. Stage 2 is performed 3 months later. The lamina is placed in the eye and the mucous membrane is placed back, allowing the cylinder to protrude through a central opening made in the mucosa. The placement of the Ahmed glaucoma drainage device can be performed along with or following any of the above-mentioned stages. DEFINITION OF GLAUCOMA:
Glaucoma was diagnosed based on 1 or more of the following factors: a vertical cup-to-disc ratio of 0.7 or more and/or any notching noted during stage 1A; the preoperative digital tension; and the preoperative need for use of systemic acetazolamide. Post stage 2, glaucoma was defined based on the structural appearance as mentioned above and the visual field defect corresponding to the optic nerve changes. Assessment of cup/disc asymmetry is usually not possible in these bilateral end-stage disorders, owing to nonvisualization of the disc in the other eye, and hence cannot be used as a criterion to define glaucoma. Progression was defined as an increase in the cup-to-disc ratio by 0.2 or more with an increase in the visual field defect, with raised digital tension. In eyes with advanced cupping, only the visual field changes and digital tension provided a clue toward progression.
SURGICAL TECHNIQUE:
Placement of ahmed glaucoma drainage device before or after stage 1A. A 360-degree conjunctival peritomy is performed. The superior and lateral rectus muscles are tagged to expose the superotemporal quadrant. The limbus is usually identified in most cases. The plate of the Ahmed glaucoma drainage device is placed deep in the superotemporal fornix and is
VOL. 159, NO. 3
secured 8 mm from the limbus using 8-0 nylon sutures. The tube length is adjusted to protrude3–4 mm beyond the limbus and cut accordingly in a bevel-up fashion. The anterior chamber is entered at the limbus using a 23 or 24 gauge needle and the tube is inserted into the anterior chamber through the entry site. The tube is secured midway to the sclera using 8-0 nylon sutures. The tube is covered using a scleral patch graft that is secured to the sclera using fibrin glue. The conjunctival edges are opposed to each other (in cases of severe forniceal shortening) or secured at the limbus using 8-0 vicryl sutures. Tarsorrhaphy sutures using 5-0 mersilk are placed laterally and/or medially to prevent retraction of the conjunctiva in eyes with shortened fornices. Placement of ahmed glaucoma drainage device following stage 1B. A full-thickness curvilinear incision is placed on the mucosa extending on either side of the superior rectus and the lateral rectus for 4–5 clock hours. Hemostasis is achieved with constant suction and judicious use of cautery. The buccal mucosa is dissected off the sclera and reflected back to expose the limbus. The superior and lateral rectus muscles are tagged. The superotemporal quadrant is exposed adequately for placement of the Ahmed glaucoma drainage device as described above until placement of the scleral patch graft to cover the tube. The buccal mucosa is reflected back over the scleral graft and the incised edges are sutured using 7-0 vicryl absorbable sutures. Placement of ahmed glaucoma drainage device following stage 2. A full-thickness curvilinear incision is placed on the mucosa extending from medial to the superior rectus to inferior to the lateral rectus for 4–5 clock hours. Hemostasis is achieved with constant suction and cautious use of cautery. The buccal mucosa is dissected off the sclera and reflected back to expose the lamina edge. The superior and lateral rectus muscles are tagged if required. The distance of the lamina edge from the center of the cylinder is measured and the anterior chamber entry is made flush along or just anterior to the edge of the lamina. The entry site is made within 9 mm from the center of the lamina so as to be anterior to the pars plana. The superotemporal quadrant is exposed adequately for placement of the device plate as described above until placement of the scleral patch graft to cover the tube. The presence of the tube in the eye is confirmed by visualizing the tube edge beneath the cylinder if possible. The buccal mucosa is reflected back over the scleral graft and the incised edges are sutured using 7-0 vicryl absorbable sutures (Figure 1). A schematic representation of the Ahmed glaucoma drainage device following the 3 stages is shown in Figure 2. POST–MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS FOLLOW-UP PROTOCOL: Topical lubricants and antibiotic
ointment were continued indefinitely. Systemic acetazolamide
GLAUCOMA IN MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS
483
FIGURE 1. Intraoperative pictures of Ahmed glaucoma drainage device placement following stage 2 of the modified osteo-odontokeratoprosthesis procedure. (Left) Mucosa is incised in the superotemporal quadrant close to insertion of the superior rectus and reflected off the sclera up to the edge of the osteo-odonto-alveolar lamina. The plate of the Ahmed glaucoma valve is pushed into the fornix and secured to the sclera using 8-0 nylon sutures. Distance from the center of the cylinder to entry of the tube into the eye should be less than 8–9 mm. (Middle) The end of the tube is visualized inside the eye beneath the optic. (Right) A donor scleral patch graft is placed over the tube along the length of its course, secured with fibrin glue, and the mucosa is reflected back, the edges secured with 7-0 vicryl sutures.
FIGURE 2. Schematic representation of the Ahmed glaucoma drainage device placement following various stages of the modified osteo-odonto-keratoprosthesis procedure. (Left) After stage 1A, the placement of the Ahmed glaucoma drainage device (AGDD) is similar to the placement in a routine glaucomatous eye. The differences include a scarred cornea and conjunctival scarring with shortened fornices. Lateral tarsorrhaphy (LT) performed at the end of the procedure helps prevent conjunctival retraction. (Middle) Placement of the AGDD after Stage 1BDC requires placement of mucosal incision (indicated by dotted line) in the superotemporal quadrant from the superior rectus muscle to the lateral rectus muscle. The mucosa is reflected to expose the limbus and the rest of the surgery is completed. The mucosa is sutured back using 7-0 vicryl sutures. (Right) Placement of the AGDD after stage 2 requires placement of mucosal incision in the superotemporal quadrant from the superior rectus muscle to the lateral rectus muscle. The mucosa is reflected to expose the lamina and the tube is inserted at or as close to the limbus as limited by the edge of the lamina. The point of entry into the eye should not be posterior to 8 mm from the center of the optical cylinder. The mucosa is sutured back using 70 vicryl sutures. SR, superior rectus muscle; IR, inferior rectus muscle; LR, lateral rectus muscle.
was used based on need for up to 6 months after the surgery. The eye was subjected to a thorough clinical examination, posterior segment evaluation with a 90-diopter lens, Humphrey visual field testing every 6 months, B-scan ultrasound, and spiral computed tomography of the lamina once a year. All patients had a descriptive record of the disc details at every visit. A photographic documentation at each visit was not possible. Humphrey visual fields 24-2 and 10-2 were performed in all eyes once every 6 months. 484
RESULTS DEMOGRAPHICS:
Eighty-five eyes of 82 patients underwent the MOOKP procedure at our center between March 2003 and March 2013. Of the 85 eyes that underwent the MOOKP procedure, 44 belonged to the SJS group and 37 to the chemical injury group, with a mean (range) follow-up of 36.93 months (1–90 months) and 30.42 months (1–80), respectively. Four eyes belonged to
AMERICAN JOURNAL OF OPHTHALMOLOGY
MARCH 2015
other etiologies. Twenty-one eyes of 19 patients had glaucoma as defined. One patient was referred to our center for second shunt placement surgery and was included in the study. Of these 20 patients, 15 eyes of 14 patients underwent the Ahmed glaucoma drainage device procedure (17 devices) at various stages of the MOOKP procedure. These included 10 male and 4 female patients with ages ranging from 23 to 54 years. The primary etiology was chemical injury in 9 patients and Stevens-Johnson syndrome in 5 patients. Among the SJS group, 3 had prior intraocular surgeries for corneal perforation and 1 was the patient operated elsewhere for the MOOKP procedure. The Ahmed glaucoma drainage device was implanted in the eye during stage 1A in 2 eyes, following stage 1A in 7 eyes, after stage 1Bþ1C in 1 eye, and after stage 2 in 6 eyes (twice in 1 eye, accounting for 7 valves). The other eye of the same patient had 2 devices implanted: 1following stage 1A and the other subsequent to stage 2. Clinical features, the timing of Ahmed glaucoma drainage device placement, and their outcomes have been listed in the Table.
nerve cupping during stage 1A of the procedure, the patient was keen to go ahead with the rest of the stages of MOOKP in an attempt at visual rehabilitation. However, the visual acuity did not improve beyond perception of light following the MOOKP procedure owing to the glaucomatous optic atrophy and high IOP, despite maximal systemic acetazolamide. Intraoperative findings revealed a normal anterior segment with no peripheral anterior synechiae. There was no previous record available regarding prior use of topical or systemic steroids indicating the possibility of steroid-induced glaucoma. Details of patients who underwent Ahmed glaucoma drainage devices have been tabulated in the Table. Only 2 eyes developed complications related to the Ahmed glaucoma drainage device. In 1 patient, in whom the valve procedure was done along with stage 1A, persistent hypotony occurred and required tube ligation 2 months following the procedure. This resulted in improvement in the IOP, and the rest of the stages could be performed. A best-corrected visual acuity of 20/50 was achieved. In the second eye, vitreous blocked the tube, necessitating an endoscopic vitrectomy.
OUTCOME:
Eleven of 15 eyes (73.3%) remained stable with adequate control of IOP. These 11 patients (11 eyes) continued to receive 125 mg of systemic acetazolamide 2– 3 times a day with potassium supplements. Glaucoma in 2 eyes was well controlled for 6 and 5 years following the MOOKP procedure. However, 1 eye developed endophthalmitis with subsequent phthisis and the other developed a sterile vitritis with laminar resorption, for which the keratoprosthesis was removed. The patient who had the Ahmed glaucoma drainage device placed prior to stage 2 of the procedure (done elsewhere) required a second device for raised IOP 6 months after the MOOKP and the same was placed at our center in the superonasal quadrant. The end of the earlier tube could not be visualized through the optical cylinder. The tube placed at our center developed a vitreous block for which an endoscopic vitrectomy was performed elsewhere, following which the IOP was well maintained. Retinal detachment occurred 1 year after the MOOKP procedure following yttrium-aluminum-garnet membranotomy for retroprosthetic membrane in the 1 eye. Post retinal detachment surgery, the best-corrected visual acuity was 20/400 and has been noted to be stable since then. IOP was well controlled in the remaining 7 eyes. Among the remaining 4 of the 15 eyes, deterioration was noted in the Humphrey visual field along with an increase in the digital tension over the period of follow-up in both eyes of a patient, and a second Ahmed glaucoma drainage device was implanted in the inferotemporal quadrant 6 years after the MOOKP procedure (Figure 3). The IOP was well controlled following the procedure. One patient developed a retinal detachment 3 months following the procedure. The patient was not keen for any further surgical intervention. The fourth patient was a young, 1-eyed patient of SJS and despite having noted the advanced optic
VOL. 159, NO. 3
DISCUSSION GLAUCOMA CONTINUES TO REMAIN ONE OF THE MOST
important sight-threatening concerns following any keratoprosthesis procedure.1,8 Pre-existing glaucoma could be primary or secondary to the etiology or previous intraocular surgical procedures, respectively. It can also occur de novo after the keratoprosthesis is implanted in the eye. Issues in diagnosis, monitoring, and managing glaucoma specific to eyes undergoing the MOOKP surgery are manifold. They are mainly in the IOP measurements, angle assessments, and lack of benefit of topical medications owing to the questionability of penetration of the same through the thick buccal mucosa. Systemic acetazolamide remains the only means to control the IOP when indicated.1 It may not always be possible to assess the anterior segment status by means of ultrasound biomicroscopy, especially in cases of ankyloblepharon or severe forniceal shortening. Despite these limitations, early diagnosis of and intervention for glaucoma in these eyes is of utmost importance. In the earlier long-term study by Falcinelli and associates, 66 of 181 eyes were noted to have pre-existing glaucoma and 10.4% developed de novo glaucoma after the surgery.4 In their systematic review of surgical outcomes of osteoodontokeratoprosthesis,3 Marchi and associates were noted to have reported a 33% occurrence of de novo glaucoma following MOOKP (85 eyes). The Singapore OOKP Study in 2009 reported no occurrence of glaucoma postoperatively with 5 eyes of 16 patients detected to have pre-existing glaucoma. We also noted no occurrence of glaucoma following the procedure in
GLAUCOMA IN MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS
485
486
TABLE. Details of Patients who Underwent the Ahmed Glaucoma Drainage Device Procedure at Different Stages of the Modified Osteo-odonto-keratoprosthesis Procedure
S. No
Age/Sex
Etiology
Prior Intraocular Procedures/Duration at Which AGV Done
Best BCVA
Final BCVA
Disc Statusa— Progression
Follow-up in Months
Comments
With stage 1A 1 31/M 2 24/F
Chemical Chemical
None Bulbar mucosal graft done elsewhere for ocular surface reconstruction
20/50 20/40
20/50 20/40
0.9 0.6
36 24
Hypotony—tube ligated 2 m later UBM—angle closed
After stage 1A 1 24/M 2 35/M
Chemical Chemical
PK þ KLAL 3 m after 1A 9 m after 1A
20/30 20/50
20/30 20/400
0.9 0.6
72 36
AMERICAN JOURNAL OF OPHTHALMOLOGY
7 34/F After stage 1BþC 1 54/M
Chemical
PK 2 years prior/11 m after 1A
PL
20/30
PL
63
Stable UBM—angle closed Thick RCM plastered to iris as seen during stage 1A RD following YAG Anterior segment formed on slit-lamp evaluation Diode CPC 1 m before 1A Stable UBM—angle closed High myopia—RD 3 m after stage 2— refused intervention Well-formed anterior segment noted during 1A GOA noted at 1A ?endophthalmitis-related phthisis
SJS
PK for corneal perforation 6 m prior to 1BþC/2 m after 1BþC
20/150
20/150
0.8
48
Stable
After stage 2 1 33/M
Chemical
20/30
20/60
Adv
84
20/30
20/40
Adv
72
0.9
60
Progression 6 y after 1st valve/stable 1 y after 2nd Progression 6 y after 1st valve/stable 1 y after 2nd Sterile vitritis with laminar resorption— lamina removed Stable
3
33/M
Chemical
Diode CPC 12 m after 1A
20/40
20/40
0.9
72
4 5
41/M 28/M
Chemical SJS
12 m after 1A Diode CPC – twice; 3 m after 1A
20/30 PL
20/30 PL
Adv cupping
67 36
6
33/F
SJS
1 m after 1A
PL
PL
Adv cupping
1
2
33/M
Chemical
Diode CPC/18 m after stage 26 y after stage 2 Diode CPC/6 y after stage 2
3
29/M
Chemical
3 y after stage 2
20/30
PL
4
40/F
SJS
20/80
20/120
0.7–0.9
74
5
25/M
Chemical
Stage 1A done for leaking cystoid cicatrix 3 y prior/3 y after stage 2 2 m after stage 2
20/150
20/150
0.6
48
6
50/M
SJS
2nd valve placed 6 m after stage 2
20/30
20/30
Adv cupping
24
Stable fibrous band from disc cause for reduced BCVA Stable 1st valve placed prior to stage 2— MOOKP surgery done elsewhere
MARCH 2015
Continued on next page
Adv ¼ advanced; AGV ¼ Ahmed glaucoma valve; BCVA ¼ best-corrected visual acuity; CPC ¼ cyclophotocoagulation; GOA ¼ glaucomatous optic atrophy; KLAL ¼ keratolimbal allograft; m ¼ months; PK ¼ penetrating keratoplasty; PL ¼ perception of light; RCM ¼ retrocorneal membrane; RD ¼ retinal detachment; UBM ¼ ultrasound biomicroscopy; y ¼ years; YAG ¼ yttrium-aluminum-garnet. a Disc status mentioned as cup-to-disc ratio.
Adv cupping 0.8 Disc pallor
60 12 36 48 12 0.8
PL 20/130 3/60 PL 20/400 Chemical Chemical Chemical Chemical Chemical 33/M 44/M 40/M 23/M 30/M 3 4 5 6 7
Chemical 24/M
AGV not done 1 40/M
2
Chemical
Diode CPC PK-twice; once for corneal perforation
20/60 20/130 20/400 20/30 20/400
Adv cupping NPL 20/20
5 0.6 NPL 20/30
24
Uncontrolled IOP on acetazolamide developed endophthalmitis in 5 m— removed lamina NPL owing to absolute glaucoma—same lamina shifted to the other eye Laminar resorption—removed Stable Stable Laminar resorption—removed Stable
Comments Follow-up in Months Disc Statusa— Progression Final BCVA Best BCVA Prior Intraocular Procedures/Duration at Which AGV Done Etiology Age/Sex S. No
TABLE. Details of Patients who Underwent the Ahmed Glaucoma Drainage Device Procedure at Different Stages of the Modified Osteo-odonto-keratoprosthesis Procedure (Continued )
VOL. 159, NO. 3
our series (85 eyes). Of these 85 eyes, pre-existing glaucoma was noted in 45.94% of eyes with chemical injury, in comparison to 9.1% of eyes with SJS. The finding of a pale disc in eyes following chemical injuries with a normal finger tension could be attributable to the direct toxic effect of the chemical by intraocular penetration, more commonly alkali, on the retinal ganglion cells, immediately after the injury.8 Glaucoma in SJS eyes was noted in eyes with prior corneal perforation (3 of 4 eyes, 75%; 1 operated for MOOKP elsewhere), leading to damage to both the anterior segment and occlusion of angle structures. In these eyes with pre-existing iridocorneal adhesions, a total iridodialyis is not possible and the iris has to be mechanically abscissed to the extent possible with residual angle closure. In the remaining 40 out of the 44 SJS eyes, the disc was noted to be healthy, with no de novo occurrence of glaucoma following the procedure. Stage 1A described in our series of patients is not a part of the Rome-Vienna protocol. However, performing this stage prior to harvesting the tooth allows intraoperative visualization of the optic disc and retina. This helps assess the possible visual potential of the eye as well as decide the aggressiveness with which to treat the glaucoma. The removal of the iris from its root, the lens extraction by the intracapsular technique, and the anterior vitrectomy done during this stage may play a role in reducing the intraocular pressure in certain cases. The decision to abandon further stages of the MOOKP procedure can be based on the retinal and optic nerve findings during stage 1A of the procedure rather than visualizing the posterior segment for the first time after the completion of the entire MOOKP procedure. In cases of advanced glaucomatous changes with a preoperative high digital tension despite maximum systemic acetazolamide, the Ahmed glaucoma drainage device procedure can be combined with stage 1A of MOOKP. However, combining these can lead to significant hypotony in the postoperative period, as was noted in 1 of our cases. The decision to implant the Ahmed glaucoma drainage device can therefore be taken subsequent to noting the effect of stage 1A on the IOP. Endo-cyclophotocoagulation is an alternative that can be performed during this stage after the placement of a temporary keratoprosthesis with the aid of a vitreoretinal colleague or through endoscopic visualization. If the IOP continues to remain high after stage 1A, the Ahmed glaucoma drainage device is placed prior to stage 1B. This facilitates ease of placement of the device as well as counters the IOP spike because of the mucosal graft. This is indicated if the patient is already on maximum dose of systemic acetazolamide prior to stage 1B of the procedure as it limits options for addressing the postoperative spike to only intravenous mannitol, which is effective for only a few days. Draping the buccal mucosal graft over the entire ocular surface compromises the uveoscleral outflow of aqueous, thereby leading to an increase in the IOP. In 5
GLAUCOMA IN MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS
487
FIGURE 3. 10-2 Humphrey visual field test showing progression of visual field defect between 2008 and 2013, in the (Left) right eye and (Right) left eye of the patient who underwent the modified osteo-odonto-keratoprosthesis procedure in both eyes. Following the progression noted in the Humphrey visual field, a second tube shunt was placed in the inferotemporal quadrant of both the eyes of the patient.
out of 7 eyes, IOP remained stable over the period of followup, in the group wherein the Ahmed glaucoma drainage device was placed before stage 1B. In such cases where the Ahmed glaucoma drainage device is placed prior to buccal mucosal grafting, care has to be taken during subsequent procedures to avoid inadvertent damage to the bleb or the tube while dissecting the conjunctiva or mucosa and tagging the recti muscles. A specific note of the device quadrant should be made in the operative notes. In this context it is important to cover the tube with a scleral patch graft to avoid erosion and exposure of the tube through the mucosa, to prevent inadvertent cutting of the tube during subsequent surgeries, as well as to prevent erosion of the tube by the lamina placed subsequently over it. While performing the Ahmed glaucoma drainage device procedure after stage 2, the point of entry into the eye is important. It is best to be as anterior as possible in order to avoid inadvertent entry through the retina as well as remain parallel and flush with the cornea. This anterior placement of the tube makes the possibility of vitreous blocking the tube remote. The distance from the center of the optical cylinder for tube entry should be anterior to 8–9 mm. Though an endoscopic endocyclophotocoagulation can be an alternate means to a shunt procedure, earlier reports have not been very encouraging. Kumar and associates have described the use of transconjunctival photocoagulation during stage 2 in 3 eyes with glaucoma. Endoscopic diode cyclophotocoagulation was performed in 2 eyes with progression, of which 1 developed endophthalmitis.5 The authors do not have experience with endoscopic endocyclophotocoagulation. Five out of 6 eyes that underwent Ahmed glaucoma drainage device implantation 488
after stage 2 and 6 of 7 eyes that underwent implantation after stage 1A showed advanced glaucomatous damage. Three of the 6 eyes requiring first or second shunt procedure after stage 2 showed a progression in the glaucomatous damage. Only 1 eye required a second shunt for progression when the first was performed after stage 1A. Placing the tube early in the stages of the MOOKP procedure could play a role in preventing the impact of IOP spikes on the nerve during the subsequent procedures as well as longterm progression of glaucomatous damage. There have been reports of glaucoma progression and appearance of de novo glaucoma following any keratoprosthesis procedure. The mechanism for the same following the Boston type 1 keratoprosthesis has been attributed to iridocorneal adhesions and/or inflammatory debris clogging the trabecular meshwork.9 Following the MOOKP procedure wherein a total iridodialysis and intracapsular cataract extraction is performed routinely, there exists very minimal possibility of iridocorneal adhesions forming owing to either the iris or capsular remnant. The anatomic distortion of the angle structures caused by a large-diameter corneal graft in a Boston type 1 keratoprosthesis does not occur in the MOOKP procedure. A central 3.5- to 4-mm opening based on the optical cylinder dimension is the only opening into the eye. There are no peripheral corneal sutures distorting the angle. These could be possible explanations for absence of de novo glaucoma occurring following the MOOKP procedure in our series. The perioperative spike in the intraocular pressure noted in almost all our cases occurs owing to a compromise caused in the uveoscleral outflow of the aqueous by the mucous membrane graft draped over the ocular surface up to the insertion of the recti muscles. This spike is managed by means of intravenous mannitol
AMERICAN JOURNAL OF OPHTHALMOLOGY
MARCH 2015
for 3–4 days in the immediate postoperative period followed by systemic acetazolamide. Systemic acetazolamide is usually continued for a period of 6 months to 1 year following the MOOKP procedure in eyes with no preexisting glaucoma. The means to monitoring the progression of glaucomatous changes is by measuring the digital tension, serial Humphrey visual field examinations, and optic disc photography. Since each stage of the MOOKP procedure causes a varied response on the IOP, the decision to place the shunt should be based on pre- and postprocedure assessment of IOP. The amount of overfiltration might be more controlled with the Ahmed glaucoma drainage device compared to the Baervaldt tube,5 though the authors have no experience with the Baervaldt tube in their series of patients. Tan and associates have described a Baerveldt drainage device without tube connection being placed in 1 eye for future use if required during stage 1 of the procedure.5 From the authors’ experience, it is better to place the drainage device after stage 1A, rather than concurrently, if the digital tension continues to remain high despite maximum systemic acetazolamide treatment. There is no role of topical antiglaucoma medications at this stage as there is very limited penetration through the buccal mucosa later. Mucosal grafting will invariably cause further spike in the IOP. The authors prefer to avoid intraocular
surgery along with placement of mucous membrane graft during stage 1BþC of the procedure owing to the possibility of an infection from the mucous membrane. It might be desirable to leave the intraocular length of the tube long enough to be visualized beneath the optic cylinder to enable visualization of any tube blockage as well as decrease the possibility of tube retraction and thereby failure to function. The need for a second valve (3 eyes in our series [2 eyes after 6 years following detection of progression on fields, and 1 after 6 months]) could be owing to failure of the drainage device after a few years. The location for the second device would preferably be in the inferotemporal quadrant, followed by the superonasal quadrant. Preplacement of the shunt in all cases undergoing the MOOKP might not be warranted considering the absence of de novo glaucoma in our series of patients. There have been earlier studies regarding issues in monitoring the IOP in eyes undergoing the MOOKP procedure as well as studies suggesting the role of the Ahmed glaucoma drainage device in these eyes. The authors believe that this is the first study reporting the outcome of the Ahmed glaucoma drainage device in MOOKP eyes, describing in detail the technique and timing of the same, along with highlighting the role of an additional stage 1A to the MOOKP procedure.
ALL AUTHORS HAVE COMPLETED AND SUBMITTED THE ICMJE FORM FOR DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST and none were reported. The authors indicate no funding support. Contributions of authors: concept and design of the study (G.I., B.S., V.L.); analysis and interpretation (G.I., B.S., V.L.); writing the article (G.I., B.S.); critical revision of the article (G.I., B.S., V.L.); final approval (G.I., B.S., V.L.); data collection (G.I., B.S., S.A., R.S.); provision of materials, patients, or resources (G.I., B.S., S.A., R.S., S.K., S.B., L.V.); statistical expertise (G.I.); literature search (G.I., B.S.). The authors acknowledge the contribution of Dr Parthopratim Dutta Majumder, Associate Consultant, Medical Research Foundation, towards schematic representation of the placement of the Ahmed Glaucoma Drainage Device following various stages of the modified osteo-odonto-keratoprosthesis procedure in the form of line diagrams in Figure 2.
REFERENCES 1. Hille K, Grabner G, Liu C, et al. Standards for modified osteoodontokeratoprosthesis (OOKP) surgery according to Strampelli and Falcinelli: the Rome-Vienna Protocol. Cornea 2005;24(8):895–908. 2. Liu C, Okera S, Tandon R, Herold J, Hull C, Thorp S. Visual rehabilitation in end-stage inflammatory ocular surface disease with the osteo-odonto-keratoprosthesis: results from the UK. Br J Ophthalmol 2008;92(9):1211–1217. 3. Tan A, Tan DT, Tan XW, Mehta JS. Osteo-odonto keratoprosthesis: systematic review of surgical outcomes and complication rates. Ocul Surf 2012;10(1):15–25. 4. Falcinelli GC, Falsini B, Taloni M, Piccardi M, Falcinelli G. Detection of glaucomatous damage in patients with osteoodontokeratoprosthesis. Br J Ophthalmol 1995;79(2):129–134.
VOL. 159, NO. 3
5. Kumar RS, Tan DT, Por YM, et al. Glaucoma management in patients with osteo-odonto-keratoprosthesis (OOKP): the Singapore OOKP Study. J Glaucoma 2009;18(5):354–360. 6. Lee RM, Al Raqqad N, Gomaa A, Steel DH, Bloom PA, Liu CS. Endoscopic cyclophotocoagulation in osteo-odontokeratoprosthesis (OOKP) eyes. J Glaucoma 2011;20(1):68–69. 7. Hille K, Hille A, Ruprecht KW. Medium term results in keratoprostheses with biocompatible and biological haptic. Graefes Arch Clin Exp Ophthalmol 2006;244(6): 696–704. 8. Cade F, Grosskreutz CL, Tauber A, Dohlman CH. Glaucoma in eyes with severe chemical burn, before and after keratoprosthesis. Cornea 2011;30(12):1322–1327. 9. Crnej A, Paschalis EI, Salvador-Culla B, et al. Glaucoma progression and role of glaucoma surgery in patients with Boston keratoprosthesis. Cornea 2014;33(4):349–354.
GLAUCOMA IN MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS
489
Biosketch Geetha Iyer completed her post graduation in Ophthalmology and Cornea Fellowship from and is currently working as Senior Consultant at the prestigious eye institute Sankara Nethralaya, Chennai, India. She specializes in cornea and ocular surface disorders’ management and has so far done more than 100 keratoprosthesis surgeries (MOOKP, Boston Type 1 and 2 Kpro). Her current ocular research is focused on Stevens Johnson syndrome and chemical injuries.
489.e1
AMERICAN JOURNAL OF OPHTHALMOLOGY
MARCH 2015
Biosketch Bhaskar Srinivasan completed his Cornea Fellowship from and is currently working as Senior Consultant at the prestigious eye institute Sankara Nethralaya, Chennai, India. He is a clinician specializing in cornea and ocular surface disorders including keratoprosthesis and his focused areas of interest include chemical injuries to the eye, lamellar and refractive surgeries.
VOL. 159, NO. 3
GLAUCOMA IN MODIFIED OSTEO-ODONTO-KERATOPROSTHESIS
489.e2
