Surgical Treatment of Strabismus Secondary to Glaucoma Drainage Device

Objective  To describe surgical strategies in a series of patients with diplopia following implantation of a glaucoma drainage device.

Methods  Retrospective review of 9 consecutive patients who underwent strabismus surgery because of strabismus and diplopia after implantation of a glaucoma drainage device.

Results  Seven patients with marked limitation to ocular rotations and incomitant strabismus underwent surgery on the eye with the implant. Two patients with mild limitation to ocular rotations of the involved eye underwent surgery on the contralateral eye. All patients had a large fibrous capsule surrounding the implant plate, adjacent muscles, and sclera. Intraocular pressure was not elevated postoperatively. Postoperative diplopia in the primary position was eliminated in 5 patients and markedly improved in 3 patients.

Conclusions  Strabismus following implantation of a glaucoma drainage device is an uncommon but serious complication. Restoration of ocular alignment is a complex undertaking requiring strabismus and glaucoma surgical expertise. Multiple surgical complications may occur. Surgical intervention may require complete removal of the fibrous capsule surrounding the implant and involved adjacent structures. Size reduction of the implant plate is helpful and did not interfere with postoperative intraocular pressure control in this study. Surgery on the contralateral eye is an option in patients with mild restriction.

Strabismus following implantation of a glaucoma drainage device has been reported by several authors.^1-12 In many cases, the diplopia is transient because it is caused by extraocular muscle and orbital tissue edema.¹ However, there are cases of permanent strabismus secondary to a mass effect, a posterior fixation suture effect, the incorporation of extraocular muscles in the capsule surrounding the implant, fat adherence to the implant, and/or direct trauma to the extraocular muscles.^1-7

The incidence of permanent strabismus after implantation of glaucoma drainage devices ranges from 6% to 100% and appears to depend on the type of implant used, where the implant is placed, and the implant material.¹ The bigger the implant, the higher the risk of developing motility disturbances. Implants that require positioning of the device's edges under the extraocular muscles have a higher incidence of ischemia, necrosis, direct trauma, and scarring between muscles and the implant.^2,8 Implantation of a glaucoma drainage device in the superonasal quadrant has been reported to cause pseudo-Brown syndrome.^3,8-11

Treatment of strabismus following implantation of glaucoma drainage devices is challenging. Deviations are usually incomitant. Some patients may be reluctant to consent to surgery on the contralateral eye. In addition, strabismus surgery on the eye containing the glaucoma drainage device may jeopardize the implant's function.^1,5

To our knowledge, there have been no reports describing the different surgical approaches for correcting strabismus after the implantation of glaucoma drainage devices. The purpose of this study was to describe the different surgical strategies used in a series of patients with diplopia following implantation of a glaucoma drainage device.

We conducted a retrospective review of 9 patients with strabismus and diplopia after implantation of a glaucoma drainage device. All patients had extraocular muscle surgery performed by the same surgeon (A.L.R.) between January 1, 1997, and December 31, 2005. Appropriate internal review board approval was obtained.

All patients underwent an ocular examination that included the measurement of distance acuity, near visual acuity, and cycloplegic refraction (with the use of cyclopentolate hydrochloride 1%); a slitlamp examination; and a dilated fundus examination. The visual acuity was measured for distance with best correction using a binocular testing system (Mentor Binocular Visual Acuity Testing; Medtronics Solan, Jacksonville, Florida) with Snellen optotypes at 20 feet. Intraocular pressure was measured with the Goldman applanation tonometer preoperatively and postoperatively on the first day after surgery, 2 months later, and then periodically as recommended by the glaucoma specialist. All patients' ductions and versions were assessed preoperatively and postoperatively in the 9 gaze positions (Figure 1).¹¹

Motor alignment was determined by prism cover testing while the subject was fixing at a 20/70 target at 6 meters and at 35.6 cm. A prism bar was held in front of the affected eye while an optical occluder covered the prism over the nonfixing eye. The patient was asked to fix with the unaffected eye. The optical occluder was then moved away from the affected eye to cover the unaffected eye. The angle of deviation was determined by measuring the amount of prism required to neutralize the movement of the nonfixing eye. In cases where the visual acuity in either eye was 1.3 or more logarithm of the minimum angle of resolution lines, the Krimsky test was performed to assess ocular alignment.¹³ Sensory response was evaluated with the Titmus stereo test (Stereo Optical Co, Inc, Chicago, Illinois).

Subjective torsion was measured with the double Maddox rod test. Two Maddox rod cylinders, one red and one green, were placed in a trial frame. The patient was asked to rotate each knob on the frame until the red and green lines were both parallel to the floor. The amount of cyclotorsional deviation (in degrees) was determined directly from the axis scale on the trial frame after the patient made the proper adjustment. Objective torsion was determined by the position of the macula relative to the optic nerve during examination of the posterior pole with the use of an indirect ophthalmoscope and a 20-diopter lens, as described elsewhere.¹⁴

A forced-duction test was performed at the beginning of the surgical procedure and then intraoperatively to determine the area of mechanical restriction and to confirm release of the restriction, respectively. This was performed by carefully grabbing the limbus with a toothed forceps on the opposite side of the gaze limitation, causing slight protrusion of the eye, and then rotating the eye in the direction of the limited duction.¹⁵ In all cases, the origin of restriction was determined to be in the area of the valve.

The ultimate goal was to free all restrictions. To achieve this, we removed (as much as technically possible) the fibrous capsule surrounding the valve and its extensions (Figure 2 and Figure 3). A limbal incision was performed to isolate the superior rectus and lateral rectus muscles. In dissecting these muscles, we encountered the fibrous capsule adjacent to or extending along the bellies of the superior rectus or lateral rectus muscles. It then was necessary to excise the fibrous capsule to reveal the muscle and free the restriction.

The preoperative and operative patient characteristics are summarized in Table 1. The mean age at the time of strabismus surgery was 71.2 years (age range, 49-86 years). Eight of the 9 patients reported diplopia. Six of those 8 patients (75%) had vertical and horizontal diplopia, 1 (12%) had vertical diplopia, and 1 (12%) had vertical, horizontal, and torsional diplopia. The mean interval between implantation of the glaucoma drainage device and the onset of diplopia was 23.5 days (range, 1-90 days). No patient had diplopia before the glaucoma drainage device was implanted.

Six of 9 patients (67%) had exotropia and vertical deviation in the primary position. Four of those 6 patients (67%) had hypotropia of the affected eye and 2 of the 6 patients (33%) had hypertropia of the affected eye. Two of the 9 patients (22%) had esotropia and vertical deviation in the primary position.

The mean preoperative intraocular pressure was 12.7 mm Hg preoperatively (range, 6-16 mm Hg). Eight patients had 1 implant in 1 eye and 1 patient had 2 implants in 1 eye. Seven devices were implanted in the superotemporal quadrant and 3 were implanted in the superonasal quadrant.

Intraoperative forced-duction test results showed restriction to ocular rotations in all 9 patients. Four patients had restriction to ocular rotations in all fields. Two of the 3 patients with superonasally implanted devices had restriction to adduction and elevation in adduction. One patient with a superotemporal implant had restriction to downgaze and adduction.

The surgical approach was individually tailored to each patient. All of the surgical procedures, except for 1 reoperation, were performed with the patient under general anesthesia. All patients underwent limbal conjunctival incision near the drainage device. The incision was extended with 2 radial incisions to the area of the adjacent rectus extraocular muscle for better exposure of the muscle, drainage device, and scar tissue. All patients underwent strabismus surgery with an adjustable suture technique; sutures were adjusted on the first postoperative day. At the end of the procedure, 0.5 mL of dexamethasone was injected subconjunctivally in the area adjacent to the extraocular muscle undergoing operation.

Seven patients underwent surgery on the eye with the implant. All of them had marked limitation to ocular rotations and incomitant strabismus. Two patients with mild limitation to ocular rotations of the involved eye underwent strabismus surgery on the contralateral eye.

All patients had a large fibrous capsule surrounding the implant plate, adjacent muscles, and sclera. Once the capsule was trimmed and opened, aqueous fluid was discharged. To prevent a flat chamber and complications such as cataract, severe hypotony, or decentered intraocular lenses, a clear corneal paracentesis was performed by the glaucoma specialist (A.L.C.) and the anterior chamber was reformed by injecting either sodium hyaluronate (Healon; Advanced Medical Optics Uppsala AB, Uppsala, Sweden) or balanced saline solution through the paracentesis. In all of the patients, scar tissue incorporated the adjacent extraocular muscles.

In 2 patients, the implant was trimmed and repositioned, which eliminated the mass effect of the implant. Three patients with superonasal implants had scar tissue between the superior oblique tendon and the superior rectus muscle. In these 3 patients, the scar tissue was removed and the superior oblique tendon was repositioned. In 1 patient the anterior portion of the superior oblique tendon was disinserted from the sclera and was attached to the surrounding scar tissue. Another patient with a functional Baerveldt implant had a dense capsule between the superior and lateral rectus muscles that extended 21 mm posteriorly to the limbus. It was necessary to remove the Baerveldt implant to expose all the scar tissue attachments and to release the restriction. The Baerveldt implant was replaced with an Ahmed glaucoma drainage device, which was implanted inferonasally to avoid the superotemporal quadrant.

Postoperative follow-up ranged from 1 to 48 months (Table 2). At the last postoperative follow-up visit, the mean intraocular pressure was 13.0 mm Hg (range, 9-18 mm Hg). After surgery, no patient had uncontrolled intraocular pressure on any examination. Two patients with preoperative diplopia and 1 patient without preoperative diplopia owing to poor vision were diplopia free in all fields of gaze after the initial strabismus surgery. Two patients were diplopia free in the primary position after they underwent a second strabismus surgery because of an inferior rectus muscle slippage. Diplopia in the primary position improved but persisted in 3 patients: 2 patients had a smaller deviation and were diplopia free with prisms and 1 had intermittent diplopia.

The incidence of strabismus following implantation of glaucoma drainage devices is variable; it appears to be more common following placement of large-plate implants. Previous authors^1,14,16,17 have reported more than an 80% incidence of strabismus following implantation of the 350-mm² Baerveldt implant or the Krupin valve with disc. In our study, 7 patients had had an Ahmed valve implanted. To our knowledge, the incidence of strabismus following Ahmed valve implantation is not reported in the literature, and there are only a few reports of strabismus after implantation of the device.^6,18 The Ahmed valve may induce fewer motility disturbances because its surface area (184 mm²) is smaller than that of the double-plate Molteno (270 mm²) or the Baerveldt (350 mm²) implant.^1,19

We found that limitation to ocular rotations in fields of gaze away from the implant and deviation toward the implant is most likely due to the presence of the fibrous capsule surrounding the implant, which causes marked restriction to ocular rotations. Superonasal implantation may create hypotropia with limitation to elevation in adduction.^2,6,9 Two of our patients had limitation to elevation in adduction; both patients had implants in the superonasal quadrant.

One patient had 13° of excyclotorsion. Surgical exploration showed partial disinsertion of the anterior half of the superior oblique tendon, which was sitting directly under the implanted valve. As soon as the valve was elevated, prior to any tissue dissection, it was apparent that the tendon was not in its normal location. The anterior portion of the superior oblique tendon most likely was accidentally detached from its insertion during implantation of the valve. The superior oblique tendon was reinserted in its normal position with a nonabsorbable suture, resulting in the absence of excyclotorsion postoperatively (Figure 4).

We found it necessary to eliminate the restriction caused by the fibrous capsule that surrounds the implant in 5 of our 9 patients. We also found it necessary to explore the extraocular muscles adjacent to the implant in all cases to ensure that they were intact and to excise the fibrous capsule. This was contrary to previous reports that recommended only recession of the adjacent or opposing muscles to avoid creating a persistent restriction.¹

The fibrous capsule was almost impossible to dissect in its entirety. The scar tissue, which was extensive in all cases, was removed until the restriction on forced-duction testing was released or until it was no longer technically possible to continue. The fibrous capsule will probably recur once the implant begins to function because of the continuous aqueous flow. However, the smaller size of the drainage device (accomplished by trimming the valve or replacing it with a smaller device), extensive dissection, intraoperative corticosteroid injection under the Tenon capsule, and the fact that the implant was situated as far as possible from the adjacent extraocular muscles may prevent recurrence of the restriction.

The main concern when operating on the eye that has the drainage implant is the risk of damaging the implant's ability to control intraocular pressure.^1,2,19 These patients usually have advanced glaucoma and intraocular pressure that is difficult to control. It is advantageous to have a glaucoma surgeon available to help with managing changes in aqueous flow when the fibrous capsule is opened, trimming or replacing the glaucoma drainage implant, and reforming the anterior chamber. When the integrity of the fibrous capsule is violated, aqueous humor is immediately discharged and the anterior chamber flattens. This is the primary reason for including a glaucoma specialist, who can immediately reform the anterior chamber with sodium hyaluronate or balanced saline solution and maintain intraocular pressure. Once the anterior chamber is reformed and the pressure is normalized, strabismus surgery can continue, restrictions can be relieved, and the appropriate muscles can be recessed. Then, the glaucoma specialist can address the issues of valve reduction and removal of viscoelastic material. Extensive scar tissue dissection may result in damage to the conjunctiva. Careful conjunctival dissection is warranted to prevent complications during conjunctival closure and to decrease the potential risk of endophthalmitis in a patient who has a glaucoma drainage device.

Strabismus surgery secondary to a glaucoma drainage device is a complex procedure with potential complications such as cataract, decentered intraocular lens, hypotony, difficulties with conjunctival closure, postoperative infection, endophthalmitis, and intraocular pressure spikes owing to retained viscoelastic material. The postoperative intraocular pressures in our patients are shown in Table 2. Postoperative intraocular pressure follow-up ranged from 1 to 48 months. No patient required additional medication or glaucoma surgery. However, this study was limited by the small number of patients and short follow-up. Longer postoperative follow-up with a glaucoma specialist is required to monitor changes in intraocular pressure and in the optic nerve.

Treatment of strabismus following the implantation of a glaucoma drainage device is challenging. Nonsurgical treatment is limited. Prisms usually do not help because the deviations are incomitant and sometimes large. To our knowledge, there are no reports of the use of chemodenervation in these patients.¹ In cases of mild restriction of the ocular rotations and a comitant strabismus deviation, surgery on the contralateral eye has the advantage of being technically easier, more predictable, and associated with less risk of damaging the glaucoma drainage implant.

From this study, we cannot conclude which characteristics of the deviation predict the need for capsule excision. Almost every case required some degree of scar tissue removal, especially when the forced-duction test result was positive. Because the fibrous capsule is very extensive, the location of the drainage device should be explored directly if the restriction needs to be released.

In summary, strabismus after implantation of a glaucoma drainage device is an uncommon but serious complication. Removal of the fibrous capsule surrounding the implant is essential to release restrictions identified on forced-duction testing. However, appropriate rectus extraocular muscle recession with the use of adjustable sutures is required to achieve ocular alignment.

Correspondence: Arthur L. Rosenbaum, MD, Department of Ophthalmology, Jules Stein Eye Institute, 100 Stein Plaza, University of California, Los Angeles, Los Angeles, CA 90095 (rosenbaum@jsei.ucla.edu).

Submitted for Publication: April 19, 2007; final revision received July 26, 2007; accepted August 5, 2007.

Financial Disclosure: None reported.

Funding/Support: Dr Rosenbaum is a recipient of a Research to Prevent Blindness Physician Scientist Merit Award.

Previous Presentation: This study was presented at the Annual Meeting of the American Association for Pediatric Ophthalmology and Strabismus; April 14, 2007; Seattle, Washington.