- Email: [email protected]
Outcome of Baerveldt glaucoma drainage implants for the treatment of uveitic glaucoma
Outcome of Baerveldt Glaucoma Drainage Implants for the Treatment of Uveitic Glaucoma Elizenda M. Ceballos, MD, Richard K. Parrish II, MD, Joyce C. Schiffman, MS Objective: To evaluate the efficacy and safety of Baerveldt glaucoma drainage devices in the management of uveitic glaucoma. Design: Retrospective, noncomparative case series. Participants: Twenty-four eyes of 24 patients who underwent implantation of Baerveldt glaucoma drainage devices between 1996 and 2000 for the treatment of uveitic glaucoma refractory to medical therapy. Intervention: Implantation of Baerveldt glaucoma drainage device. Main Outcome Measures: Control of intraocular pressure (IOP), number of glaucoma medications needed for adequate IOP control, visual acuity, complications associated with the surgery, and the effect of subsequent surgery on the ability of the device to control IOP. Success was defined as IOP ⱖ5 and ⱕ21 mmHg with or without antiglaucoma medications and without need for further glaucoma surgery, loss of light perception, or phthisis. Results: Cumulative life-table success rates were 95.8% at 3 months and 91.7% at 6 months, 12 months, and 24 months. The mean postoperative follow-up was 20.8 months. The IOP was reduced from a preoperative mean of 30.5 ⫾ 8.96 mmHg with 3.1 ⫾ 0.99 antiglaucoma medications to a postoperative mean at 6 months or 1 year of 13.0 ⫾ 4.6 mmHg (P ⬍ 0.001) with 0.8 ⫾ 0.8 antiglaucoma medications (P ⬍ 0.001). At last follow-up 14 of 24 eyes (58.3%) required no antiglaucoma medications. Best-corrected visual acuity improved or remained within 2 lines of preoperative visual acuity in 19 (79.2%) eyes. The most common complications were choroidal effusions in four (16.7%), hypotony in three (12.5%) eyes, cystoid macular edema in three (12.5%) eyes, and failure of corneal grafts in two (8.3%). Seven of 22 eyes (31.8%) in which successful control of IOP with the Baerveldt implant was achieved underwent subsequent nonglaucoma-related incisional surgery. None of these eyes (0%) lost IOP control after the subsequent procedure. Conclusions: The Baerveldt glaucoma drainage device offers reasonable safety and effectiveness for the control of IOP in eyes with uveitis and refractory glaucoma. Ophthalmology 2002;109:2256 –2260 © 2002 by the American Academy of Ophthalmology.
The management of uveitic glaucoma that is refractory to medical therapy presents a challenge to the ophthalmic surgeon. The success rate of trabeculectomy with adjunctive antiscarring agents in patients with uveitis is highly variable (50%–100%),1–5 with higher success rates correlating with shorter postoperative follow-up. Therefore, the long-term role of conventional filtering surgery in this patient population remains undetermined. In addition, young patient age, failure of previous filtering surgery, conjunctival scarring, or the anticipation for future nonglaucoma-related ocular
Originally received: August 20, 2001. Accepted: February 26, 2002. Manuscript no. 210646. From the Department of Ophthalmology, University of Miami School of Medicine, Bascom Palmer Eye Institute, Miami, Florida. Supported in part by the Heed Ophthalmic Foundation, Cleveland, Ohio. The authors have no commercial, proprietary, or financial interest in Baerveldt glaucoma drainage devices. Correspondence to Richard K. Parrish, MD, Department of Ophthalmology, 900 NW 17th St., Miami, Florida 33136.
2256
© 2002 by the American Academy of Ophthalmology Published by Elsevier Science Inc.
surgery renders trabeculectomy a suboptimal option for some patients. Few reports in the literature address the use of glaucoma drainage devices in patients with uveitic glaucoma. Da Mata et al6 reported a 1-year 94% success rate in intraocular pressure (IOP) control (IOP ⱕ21 mmHg) in 21 study eyes treated with Ahmed valve (Ahmed drainage device, New World Medical, Inc., Rancho Cucamonga, CA) implantation. Valimaki and colleagues7 reported 90% success (IOP ⱕ22 mmHg) with Molteno device (IOP, Inc., Costa Mesa, CA) implantation after 52 months of follow-up in 27 eyes of 19 patients with secondary glaucoma caused by juvenile rheumatoid arthritis. In a prospective case series, Molteno et al8 found that the probability of the Molteno implant to control IOP (IOP ⱕ21 mmHg) was 87% and 77% at 5 and 10 years after surgery, respectively. The Baerveldt glaucoma drainage device (Pharmacia & Upjohn, Kalamazoo, MI) consists of a nonvalved silicone tube that drains aqueous humor from either the anterior chamber or the vitreous cavity (in vitrectomized eyes) onto the outer surface of a convex, flexible, silicone plate. The ISSN 0161-6420/02/$–see front matter PII S0161-6420(02)01294-0
Ceballos et al 䡠 Baerveldt Implants for Uveitic Glaucoma plate is secured to the sclera over the equator of the eye with sutures and is covered with Tenon’s fascia and conjunctiva. Within weeks, a fibrous capsule envelops the plate, forming a filtering bleb. To our knowledge, no previously published reports have evaluated the safety and efficacy of Baerveldt glaucoma drainage implants in patients with uveitis. In this study, we report our experience with this drainage device in 24 patients.
Material and Methods The study population was composed of 24 eyes of 24 patients with uveitic glaucoma who underwent placement of a Baerveldt glaucoma drainage device at the Bascom Palmer Eye Institute from 1996 to 2000 and for whom at least 6 months of follow-up data are available. The patients were identified through a retrospective review of the Institute’s computerized database (Institutional Review Board research protocol #00/491). In patients who underwent bilateral Baerveldt implantation, only the first eye to undergo surgery was included in the study. Patients were selected for Baerveldt implantation on the basis of intractable uveitic glaucoma uncontrolled with maximally tolerated medical therapy. An attempt was made to maximally control the uveitis preoperatively in all patients undergoing nonurgent surgery. Topical steroids were the most commonly used antiinflammatory agents. Periocular injections of triamcinolone acetonide and oral prednisone were only rarely used perioperatively. After written informed consent was obtained, the surgery was performed by one of five glaucoma surgeons using a similar operative technique. A 5- to 6-clock hour conjunctival peritomy was created, centered on the quadrant chosen for implantation (usually the superotemporal quadrant). Dissection was carried out between Tenon’s fascia and sclera. The two rectus muscles in the quadrant chosen for implantation were isolated on muscle hooks, and the wings of the Baerveldt plate were slipped under the muscle bellies under direct visualization. The plate was secured to the sclera 10 mm posterior to the limbus with two 9-0 nylon or 9-0 polypropylene interrupted sutures placed through the plate fixation holes. The silicone tube was then tied off near the junction of the tube and plate in a watertight fashion using 7-0 polyglactin 910 suture. For pars plana placement in vitrectomized eyes, the tube was trimmed in a beveled fashion to have an intraocular segment of 4 mm and was inserted through the pars plana through the tract of a V Lance knife (Alcon Laboratories, Fort Worth, TX) blade. The tube was secured to its entry site with a 9-0 polypropylene mattress suture. For anterior chamber placement, the tube was trimmed in a beveled fashion to extend 2 to 4 mm beyond the limbus and was inserted into the anterior chamber parallel to the iris plane through the tract of a 23-gauge needle. Fenestrations were placed anterior to the polyglactin suture at the surgeon’s discretion. The tube and its insertion site were covered with a patch of glycerin-preserved donor cornea or sclera, which was sutured to the patient’s sclera with either 7-0 polyglactin 910 or 10-0 nylon sutures at the corners. The conjunctiva was reopposed to the limbus, covering the plate and patch graft, using 7-0 polyglactin 910 or 10-0 nylon sutures. A subconjunctival injection of dexamethasone sodium phosphate was given in the inferior fornix. Postoperative therapy was dependent on surgeon preference but generally consisted of a topical antibiotic four times daily for 5 to 7 days and 1% prednisolone acetate drops 4 to 12 times daily, with the initial dosing and subsequent tapering schedule based on the level of postoperative inflammation. Baerveldt plates of surface area 250 mm2, 350 mm2, and 425 mm2 were implanted in 2, 21, and 1 eyes, respectively. Twenty-
Table 1. Characteristics of Uveitic Glaucoma Patients Treated with Baerveldt Glaucoma Drainage Implants 24 Eyes, 24 Patients
Number (%)
Age (mean) Gender (number, %) Male Female Race (number, %) Hispanic Black White Asian Previous ocular surgery (number, %) No Yes Mean number of procedures Number of failed trabeculectomies Number of failed GDDs Lens status (number, %) Phakic Pseudophakic Aphakic Uveitis diagnoses (number, %) Idiopathic Herpetic Fuchs heterochromic iridochyclitis Sarcoidosis Toxoplasmosis Phacoanaphylaxis Juvenile rheumatoid arthritis HLA B27-related
57.5 yrs (range, 18–88 yrs) 8 (33.3%) 16 (66.7%) 9 (37.5%) 7 (29.2%) 7 (29.2%) 1 (4.2%) 5 (20.8%) 19 (79.2%) 1.7 8 2 5 (20.8%) 12 (50.0%) 7 (29.2%) 7 (29.2%) 6 (25.0%) 4 (16.7%) 3 (12.5%) 1 (4.2%) 1 (4.2%) 1 (4.2%) 1 (4.2%)
GDDs ⫽ glaucoma drainage devices.
two tubes were inserted into the anterior chamber, and two were inserted into the vitreous cavity. Concurrent surgeries included anterior vitrectomy with placement of an anterior chamber intraocular lens (one eye), closure of a leaking filter (one eye), pars plana vitrectomy (two eyes), and penetrating keratoplasty (one eye). The following data were collected for each eye preoperatively and at 3, 6, 9, 12, 18, 24, 30, 36, and 48 months postoperatively (when available): visual acuity, intraocular pressure, and number of antiglaucoma medications. Postoperative complications and surgical interventions during the follow-up period were also recorded. Success was defined as IOP ⱖ5 mmHg and ⱕ21 mmHg with or without antiglaucoma medications and without further glaucoma surgery and at least light perception vision. Failure was defined as phthisis, loss of light perception, need for cyclodestruction or placement of a second glaucoma drainage device, IOP ⬎ 21 mmHg with or without antiglaucoma medications on two consecutive visits, or chronic hypotony (IOP ⬍ 5 mmHg on two consecutive visits with signs or symptoms of hypotony such as choroidal effusions, hypotony maculopathy, corneal folds, and decreased vision). Paired t tests were used to compare preoperative and postoperative IOP and number of glaucoma medications. Kaplan-Meier survival analysis was used to provide success rates.
Results Twenty-four eyes of 24 patients with an average age of 57.5 years (range, 18 – 88 years) were included in the study. The patient
2257
Ophthalmology Volume 109, Number 12, December 2002
Figure 1. Kaplan-Meier life-table analysis. The cumulative probability of success after Baerveldt glaucoma drainage device implantation was 95.8% (95% confidence interval, 87.8%–99%) at 3 months and 91.7% (95% confidence interval, 80.6%–99%) at 6 months, 12 months, and thereafter.
characteristics and uveitic diagnoses are summarized in Table 1. As expected, the patients had complicated ocular histories, with 19 patients (79.2%) having had previous incisional surgeries (mean, 1.7 surgeries; range 0 – 6). Five patients (20.8%) were phakic, 12 (50%) were pseudophakic, and 7 (29.2%) were aphakic. Ten patients had failed previous glaucoma surgery: 8 patients had undergone 14 trabeculectomies (8 with known antimetabolite use), and 2 patients had undergone placement of aqueous drainage devices (one Ahmed valve and one Baerveldt implant of unknown surface area). Two surgeries failed, one at 3 months and one at 6 months postoperatively, giving a cumulative probability of success after Baerveldt glaucoma drainage device implantation of 95.8% (95% confidence interval, 87.8%–99%) at 3 months, and 91.7% (95% confidence interval, 80.6%–99%) at 6 months, 12 months, and thereafter (Fig 1). The number of evaluable patients at 9, 12, 18, and 24 months was 20, 15, 10, and 8, respectively. Both surgical failures were a result of chronic hypotony. One of the two patients developed kissing choroidal effusions with resulting macular ischemia. The choroidal effusions were drained twice but recurred, necessitating permanent ligature of the tube with a nonabsorbable suture. The mean IOPs before and after tube ligation were 2.2 mmHg and 11.2 mmHg, respectively. The second patient had chronic hypotony (mean IOP, 4.3 mmHg; range, 3.5–5.5 mmHg) with hypotony maculopathy, cystoid macular edema, and papillitis. No eyes failed because of failure of the device to control elevated IOP, and no eye lost light perception or became phthisical.
Figure 2. Preoperative and postoperative intraocular pressure (IOP) and antiglaucoma medication summary. The bar graph (left y-axis) represents the mean IOP (mmHg). The line graph (right y-axis) represents the mean number of antiglaucoma medications.
The IOP was reduced from a preoperative mean (⫾ standard deviation) of 30.5 ⫾ 8.96 mmHg to a postoperative mean at 1 year (6 months if ⬍1 year follow-up) of 13.0 ⫾ 4.6 mmHg (P ⬍ 0.001). The mean number of antiglaucoma medications needed for IOP control (⫾ standard deviation) was also reduced from a preoperative mean of 3.1 ⫾ 0.99 to 0.8 ⫾ 0.8 at 1 year (6 months if ⬍1 year follow-up, P ⬍ 0.001). Fourteen patients (58.3%) did not require any antiglaucoma medications at last follow-up. The IOP and medication summary are given in Table 2. The mean IOP and mean number of medications needed at specific time intervals after surgery are shown in Figure 2. Five patients had intraocular pressures of ⬍21 mmHg on their preoperative visit. Two of these patients had widely fluctuating intraocular pressures, which were sometimes ⬍21 mmHg. These two patients had stable intraocular pressures ⬍21 mmHg after the Baerveldt implant and were considered to have a successful outcome in the analysis. The other three patients had intraocular pressures between 18 and 19 mmHg on two to four antiglaucoma medications preoperatively. These three patients had IOP reductions of 28% to 42% on none to one antiglaucoma medication postoperatively and were also considered to have a successful outcome in the analysis.
Table 2. Intraocular Pressure and Medication Summary of Uveitic Glaucoma Patients Treated with Baerveldt Glaucoma Drainage Implants
IOP (mmHg) (mean ⫾ SD) Antiglaucoma medications (mean ⫾ SD) Zero One Two Three Four
Preoperative
1 Year Postoperative*
Last Follow-up†
30.5 ⫾ 8.96 3.1 ⫾ 0.99 1 0 4 9 10
13.0 ⫾ 4.6 0.8 ⫾ 0.8 12 6 6 0 0
13.3 ⫾ 5.1 0.8 ⫾ 1.1 14 4 5 0 1
* Six-months postoperative if ⬍ 1 year follow-up. † The mean postoperative follow-up was 20.8 months (range, 6 –52 months). IOP ⫽ intraocular pressure; SD ⫽ standard deviation.
2258
Ceballos et al 䡠 Baerveldt Implants for Uveitic Glaucoma Table 3. Postoperative Complications of Uveitic Glaucoma Patients Treated with Baerveldt Glaucoma Drainage Implants No. of Eyes 1
Description
Surgical Intervention
1*†
Limited suprachoroidal hemorrhage, hypotony maculopathy (transient), cystoid macular edema Kissing choroidal effusions, hypotony maculopathy, macular ischemia
1 1 1 1† 1* 1
Choroidal effusion (transient) Choroidal effusion (transient), hyphema Persistent low peripheral choroidal effusion Hypotony maculopathy, cystoid macular edema Cystoid macular edema Tube occluded with vitreous twice in eye with opaque cornea
1* 1
Failed corneal graft Failed corneal graft
None Choroidal effusions drained twice; tube permanently ligated None None None Religation of tube with absorbable suture None Tube repositioning anterior vitrectomy, penetrating keratoplasty, and tube repositioning Penetrating keratoplasty None
* Denotes patients who lost 2 or more lines of best-corrected visual acuity. Denotes the two surgical failures caused by chronic hypotony.
†
The preoperative best-corrected visual acuity ranged from 20/20 to hand motions. At last follow-up, 19 (79.2%) patients had a best-corrected visual acuity that was either unchanged (5 patients), improved (10 patients), or within 1 line (4 patients) of preoperative levels. Of the five patients who lost 2 or more lines of best-corrected visual acuity, the etiologies were as follows: macular scarring (one eye), macular ischemia related to kissing choroidal effusions (one eye), cystoid macular edema (one eye), amniotic membrane covering cornea at last follow-up (one eye), and failed corneal graft (one eye). There were no intraoperative complications in this series. Ten eyes had postoperative complications (Table 3). The most common were choroidal effusions (four eyes, 16.7%), hypotony maculopathy (three eyes, 12.5%), cystoid macular edema (three eyes, 12.5%), and failure of corneal grafts (two eyes, 8.3%). One patient in whom a corneal graft failed had undergone penetrating keratoplasty 3 months before placement of the Baerveldt implant. The corneal graft remained clear for 2 years and then slowly failed. The second patient in whom a corneal graft failed had undergone penetrating keratoplasty at the same time as placement of the Baerveldt implant. Her corneal graft failed 12 months after surgery, at which time she underwent a second penetrating keratoplasty. Most complications were transient and of no visual consequence. Those complications that resulted in loss of 2 or more lines of best-corrected visual acuity included kissing choroidal effusions with resulting macular ischemia (one patient, 6 lines of bestcorrected visual acuity lost), failure of corneal graft (one patient, 3 lines of best-corrected visual acuity lost), and cystoid macular edema (one patient, 2 lines of best-corrected visual acuity lost). There were no cases of tube– cornea touch, endophthalmitis, or plate extrusion. Seven of the 22 eyes (31.8%) in which the Baerveldt glaucoma drainage implant was considered a success required subsequent nonglaucoma-related ocular surgeries: excision of conjunctival inclusion cyst (one eye); phacoemulsification with posterior chamber intraocular lens implantation (one eye); superficial keratectomy with amnionic membrane transplantation (one eye), and penetrating keratoplasty alone or in combination with cataract extraction/IOL implantation, anterior vitrectomy, or tube repositioning (four eyes). All eyes that underwent post-Baerveldt penetrating keratoplasties had either a failed corneal graft (one eye) or decompensated corneas (three eyes) preoperatively. No eye that underwent incisional surgery after placement of the Baerveldt device subsequently lost control of IOP.
Discussion The best procedure for the surgical management of uveitic glaucoma refractory to medical therapy is still a subject of considerable debate. Although some studies report reasonable success rates using trabeculectomy with antimetabolites,1–5 in clinical situations involving conjunctival scarring, young patient age, a high likelihood of future ocular surgery, or a postoperative course likely to include protracted intraocular inflammation, placement of a glaucoma drainage device may be the preferred method for achieving long-term control of IOP. Our overall surgical success rate of 91.7% at 1 and 2 years with the Baerveldt glaucoma drainage implant is encouraging, especially in our highly varied patient population with multiple risk factors for failure of conventional filtering surgery: uveitis in all patients, aphakia or pseudophakia in 79.2%, black or Hispanic race in 66.7%, age ⬍60 years in 62.5%, and previously failed filtering surgery in 33.3% of patients. In addition, all the 10 patients who had failed trabeculectomy (8 patients) or previous tube shunt placement (2 patients) achieved successful IOP control with the Baerveldt device. Interestingly, none of the Baerveldt implants failed because of an inability to control elevated IOP postoperatively, and no surgeries failed after 6 months of follow-up, despite seven of the patients undergoing subsequent nonglaucoma-related incisional surgery. We also did not observe a trend with time for higher IOP or increased need for IOP-lowering medications. Although longer follow-up is needed to detect late failures, the lack of apparent attrition of success over time is encouraging, given that long-term studies of trabeculectomy with antimetabolites show that success decreases over time.9,10 Furthermore, the ability of the implant to function well despite subsequent surgical procedures is important, because these are frequently needed to deal with the complications of chronic intraocular inflammation or the steroids used for treatment. Both of the patients in whom the surgery failed did so early in the postoperative period and because of intractable
2259
Ophthalmology Volume 109, Number 12, December 2002 hypotony. These two patients were the youngest in the study (18 and 21 years old) and also had in common female gender, aphakia, and implant size (350 mm2). One patient developed kissing choroidal effusions that had to be drained twice, macular ischemia, and profound irreversible visual loss. Her tube necessitated permanent ligature. The second patient maintains an IOP in the 3.5 to 5.5 mmHg range (mean, 4.3 mmHg), with persistent inflammation, cystoid macular edema, and hypotony maculopathy. She has elected no treatment at this time. Unfortunately, the relatively small number of patients in our study and the common use of Baerveldt implants with larger surface areas (23 of 24 were 350 mm2 or larger) do not permit speculation as to whether a Baerveldt implant with a smaller surface area or a valved glaucoma drainage implant would have prevented the hypotony in these two patients. Da Mata et al6 also reported two young females (ages 17 and 28) who developed transient (one patient) and persistent (one patient) hypotony (IOP ⬍5) in their series of 19 patients who underwent placement of an Ahmed valve for the treatment of uveitic glaucoma. However, in a study by Valimaki et al7 of 27 eyes of 19 children who underwent placement of a Molteno implant (24 [89%] single plate and 3 [11%] double plate) for the treatment of secondary glaucoma associated with juvenile rheumatoid arthritis, no eyes developed persistent hypotony (IOP ⬍ 6 mmHg). Implantation of the Baerveldt glaucoma drainage device did not seem to deleteriously affect the uveitis: 12 patients who were using topical corticosteroids or nonsteroidal antiinflammatory agents preoperatively still required them at last follow-up, 5 patients who required topical steroids or topical nonsteroidal anti-inflammatory agents preoperatively did not require them at last follow-up, 5 patients who did not require topical steroids preoperatively required them at last follow-up, and a single patient needed the systemic steroids and immunosuppressants postoperatively in addition to her usual preoperative topical steroid regimen. Overall, the favorable results of this study indicate that implan-
2260
tation of the Baerveldt glaucoma drainage device is relatively safe and effective in eyes with uveitic glaucoma either as the initial surgical procedure for control of IOP or as an additional treatment modality for eyes that have failed previous glaucoma surgery.
References 1. Jampel HD, Jabs DA, Quigley HA. Trabeculectomy with 5-fluorouracil for adult inflammatory glaucoma. Am J Ophthalmol 1990;109:168 –73. 2. Prata JA Jr, Neves RA, Minckler DS, et al. Trabeculectomy with mitomycin C in glaucoma associated with uveitis. Ophthalmic Surg 1994;25:616 –20. 3. Wright MM, McGehee RF, Pederson JE. Intraoperative mitomycin-C for glaucoma associated with ocular inflammation. Ophthalmic Surg Lasers 1997;28:370 – 6. 4. Patitsas CJ, Rockwood EJ, Meisler DM, Lowder CY. Glaucoma filtering surgery with postoperative 5-fluorouracil in patients with intraocular inflammatory disease. Ophthalmology 1992;99:594 –9. 5. Towler HMA, Bates AK, Broadway DC, Lightman S. Primary trabeculectomy with 5-fluorouracil for glaucoma secondary to uveitis. Ocul Immunol Inflamm 1995;3:163–70. 6. Da Mata A, Burk SE, Netland PA, et al. Management of uveitic glaucoma with Ahmed glaucoma valve implantation. Ophthalmology 1999;106:2168 –72. 7. Valimaki J, Airaksinen JA, Tuulonen A. Molteno implantation for secondary glaucoma in juvenile rheumatoid arthritis. Arch Ophthalmol 1997;115:1253– 6. 8. Molteno ACB, Sayawat N, Herbison P. Otago glaucoma surgery outcome study: long-term results of uveitis with secondary glaucoma drained by Molteno implants. Ophthalmology 2001;108:605–13. 9. Araie M, Shoji N, Shirato S, Nahano Y. Postoperative subconjunctival 5-fluorouracil injections and success probability of trabeculectomy in Japanese: results of 5-year follow-up. Jpn J Ophthalmol 1992;36:158 – 68. 10. Three-year follow-up of the Fluorouracil Filtering Surgery Study. Am J Ophthalmol 1993;115:82–92.
The management of uveitic glaucoma that is refractory to medical therapy presents a challenge to the ophthalmic surgeon. The success rate of trabeculectomy with adjunctive antiscarring agents in patients with uveitis is highly variable (50%–100%),1–5 with higher success rates correlating with shorter postoperative follow-up. Therefore, the long-term role of conventional filtering surgery in this patient population remains undetermined. In addition, young patient age, failure of previous filtering surgery, conjunctival scarring, or the anticipation for future nonglaucoma-related ocular
Originally received: August 20, 2001. Accepted: February 26, 2002. Manuscript no. 210646. From the Department of Ophthalmology, University of Miami School of Medicine, Bascom Palmer Eye Institute, Miami, Florida. Supported in part by the Heed Ophthalmic Foundation, Cleveland, Ohio. The authors have no commercial, proprietary, or financial interest in Baerveldt glaucoma drainage devices. Correspondence to Richard K. Parrish, MD, Department of Ophthalmology, 900 NW 17th St., Miami, Florida 33136.
2256
© 2002 by the American Academy of Ophthalmology Published by Elsevier Science Inc.
surgery renders trabeculectomy a suboptimal option for some patients. Few reports in the literature address the use of glaucoma drainage devices in patients with uveitic glaucoma. Da Mata et al6 reported a 1-year 94% success rate in intraocular pressure (IOP) control (IOP ⱕ21 mmHg) in 21 study eyes treated with Ahmed valve (Ahmed drainage device, New World Medical, Inc., Rancho Cucamonga, CA) implantation. Valimaki and colleagues7 reported 90% success (IOP ⱕ22 mmHg) with Molteno device (IOP, Inc., Costa Mesa, CA) implantation after 52 months of follow-up in 27 eyes of 19 patients with secondary glaucoma caused by juvenile rheumatoid arthritis. In a prospective case series, Molteno et al8 found that the probability of the Molteno implant to control IOP (IOP ⱕ21 mmHg) was 87% and 77% at 5 and 10 years after surgery, respectively. The Baerveldt glaucoma drainage device (Pharmacia & Upjohn, Kalamazoo, MI) consists of a nonvalved silicone tube that drains aqueous humor from either the anterior chamber or the vitreous cavity (in vitrectomized eyes) onto the outer surface of a convex, flexible, silicone plate. The ISSN 0161-6420/02/$–see front matter PII S0161-6420(02)01294-0
Ceballos et al 䡠 Baerveldt Implants for Uveitic Glaucoma plate is secured to the sclera over the equator of the eye with sutures and is covered with Tenon’s fascia and conjunctiva. Within weeks, a fibrous capsule envelops the plate, forming a filtering bleb. To our knowledge, no previously published reports have evaluated the safety and efficacy of Baerveldt glaucoma drainage implants in patients with uveitis. In this study, we report our experience with this drainage device in 24 patients.
Material and Methods The study population was composed of 24 eyes of 24 patients with uveitic glaucoma who underwent placement of a Baerveldt glaucoma drainage device at the Bascom Palmer Eye Institute from 1996 to 2000 and for whom at least 6 months of follow-up data are available. The patients were identified through a retrospective review of the Institute’s computerized database (Institutional Review Board research protocol #00/491). In patients who underwent bilateral Baerveldt implantation, only the first eye to undergo surgery was included in the study. Patients were selected for Baerveldt implantation on the basis of intractable uveitic glaucoma uncontrolled with maximally tolerated medical therapy. An attempt was made to maximally control the uveitis preoperatively in all patients undergoing nonurgent surgery. Topical steroids were the most commonly used antiinflammatory agents. Periocular injections of triamcinolone acetonide and oral prednisone were only rarely used perioperatively. After written informed consent was obtained, the surgery was performed by one of five glaucoma surgeons using a similar operative technique. A 5- to 6-clock hour conjunctival peritomy was created, centered on the quadrant chosen for implantation (usually the superotemporal quadrant). Dissection was carried out between Tenon’s fascia and sclera. The two rectus muscles in the quadrant chosen for implantation were isolated on muscle hooks, and the wings of the Baerveldt plate were slipped under the muscle bellies under direct visualization. The plate was secured to the sclera 10 mm posterior to the limbus with two 9-0 nylon or 9-0 polypropylene interrupted sutures placed through the plate fixation holes. The silicone tube was then tied off near the junction of the tube and plate in a watertight fashion using 7-0 polyglactin 910 suture. For pars plana placement in vitrectomized eyes, the tube was trimmed in a beveled fashion to have an intraocular segment of 4 mm and was inserted through the pars plana through the tract of a V Lance knife (Alcon Laboratories, Fort Worth, TX) blade. The tube was secured to its entry site with a 9-0 polypropylene mattress suture. For anterior chamber placement, the tube was trimmed in a beveled fashion to extend 2 to 4 mm beyond the limbus and was inserted into the anterior chamber parallel to the iris plane through the tract of a 23-gauge needle. Fenestrations were placed anterior to the polyglactin suture at the surgeon’s discretion. The tube and its insertion site were covered with a patch of glycerin-preserved donor cornea or sclera, which was sutured to the patient’s sclera with either 7-0 polyglactin 910 or 10-0 nylon sutures at the corners. The conjunctiva was reopposed to the limbus, covering the plate and patch graft, using 7-0 polyglactin 910 or 10-0 nylon sutures. A subconjunctival injection of dexamethasone sodium phosphate was given in the inferior fornix. Postoperative therapy was dependent on surgeon preference but generally consisted of a topical antibiotic four times daily for 5 to 7 days and 1% prednisolone acetate drops 4 to 12 times daily, with the initial dosing and subsequent tapering schedule based on the level of postoperative inflammation. Baerveldt plates of surface area 250 mm2, 350 mm2, and 425 mm2 were implanted in 2, 21, and 1 eyes, respectively. Twenty-
Table 1. Characteristics of Uveitic Glaucoma Patients Treated with Baerveldt Glaucoma Drainage Implants 24 Eyes, 24 Patients
Number (%)
Age (mean) Gender (number, %) Male Female Race (number, %) Hispanic Black White Asian Previous ocular surgery (number, %) No Yes Mean number of procedures Number of failed trabeculectomies Number of failed GDDs Lens status (number, %) Phakic Pseudophakic Aphakic Uveitis diagnoses (number, %) Idiopathic Herpetic Fuchs heterochromic iridochyclitis Sarcoidosis Toxoplasmosis Phacoanaphylaxis Juvenile rheumatoid arthritis HLA B27-related
57.5 yrs (range, 18–88 yrs) 8 (33.3%) 16 (66.7%) 9 (37.5%) 7 (29.2%) 7 (29.2%) 1 (4.2%) 5 (20.8%) 19 (79.2%) 1.7 8 2 5 (20.8%) 12 (50.0%) 7 (29.2%) 7 (29.2%) 6 (25.0%) 4 (16.7%) 3 (12.5%) 1 (4.2%) 1 (4.2%) 1 (4.2%) 1 (4.2%)
GDDs ⫽ glaucoma drainage devices.
two tubes were inserted into the anterior chamber, and two were inserted into the vitreous cavity. Concurrent surgeries included anterior vitrectomy with placement of an anterior chamber intraocular lens (one eye), closure of a leaking filter (one eye), pars plana vitrectomy (two eyes), and penetrating keratoplasty (one eye). The following data were collected for each eye preoperatively and at 3, 6, 9, 12, 18, 24, 30, 36, and 48 months postoperatively (when available): visual acuity, intraocular pressure, and number of antiglaucoma medications. Postoperative complications and surgical interventions during the follow-up period were also recorded. Success was defined as IOP ⱖ5 mmHg and ⱕ21 mmHg with or without antiglaucoma medications and without further glaucoma surgery and at least light perception vision. Failure was defined as phthisis, loss of light perception, need for cyclodestruction or placement of a second glaucoma drainage device, IOP ⬎ 21 mmHg with or without antiglaucoma medications on two consecutive visits, or chronic hypotony (IOP ⬍ 5 mmHg on two consecutive visits with signs or symptoms of hypotony such as choroidal effusions, hypotony maculopathy, corneal folds, and decreased vision). Paired t tests were used to compare preoperative and postoperative IOP and number of glaucoma medications. Kaplan-Meier survival analysis was used to provide success rates.
Results Twenty-four eyes of 24 patients with an average age of 57.5 years (range, 18 – 88 years) were included in the study. The patient
2257
Ophthalmology Volume 109, Number 12, December 2002
Figure 1. Kaplan-Meier life-table analysis. The cumulative probability of success after Baerveldt glaucoma drainage device implantation was 95.8% (95% confidence interval, 87.8%–99%) at 3 months and 91.7% (95% confidence interval, 80.6%–99%) at 6 months, 12 months, and thereafter.
characteristics and uveitic diagnoses are summarized in Table 1. As expected, the patients had complicated ocular histories, with 19 patients (79.2%) having had previous incisional surgeries (mean, 1.7 surgeries; range 0 – 6). Five patients (20.8%) were phakic, 12 (50%) were pseudophakic, and 7 (29.2%) were aphakic. Ten patients had failed previous glaucoma surgery: 8 patients had undergone 14 trabeculectomies (8 with known antimetabolite use), and 2 patients had undergone placement of aqueous drainage devices (one Ahmed valve and one Baerveldt implant of unknown surface area). Two surgeries failed, one at 3 months and one at 6 months postoperatively, giving a cumulative probability of success after Baerveldt glaucoma drainage device implantation of 95.8% (95% confidence interval, 87.8%–99%) at 3 months, and 91.7% (95% confidence interval, 80.6%–99%) at 6 months, 12 months, and thereafter (Fig 1). The number of evaluable patients at 9, 12, 18, and 24 months was 20, 15, 10, and 8, respectively. Both surgical failures were a result of chronic hypotony. One of the two patients developed kissing choroidal effusions with resulting macular ischemia. The choroidal effusions were drained twice but recurred, necessitating permanent ligature of the tube with a nonabsorbable suture. The mean IOPs before and after tube ligation were 2.2 mmHg and 11.2 mmHg, respectively. The second patient had chronic hypotony (mean IOP, 4.3 mmHg; range, 3.5–5.5 mmHg) with hypotony maculopathy, cystoid macular edema, and papillitis. No eyes failed because of failure of the device to control elevated IOP, and no eye lost light perception or became phthisical.
Figure 2. Preoperative and postoperative intraocular pressure (IOP) and antiglaucoma medication summary. The bar graph (left y-axis) represents the mean IOP (mmHg). The line graph (right y-axis) represents the mean number of antiglaucoma medications.
The IOP was reduced from a preoperative mean (⫾ standard deviation) of 30.5 ⫾ 8.96 mmHg to a postoperative mean at 1 year (6 months if ⬍1 year follow-up) of 13.0 ⫾ 4.6 mmHg (P ⬍ 0.001). The mean number of antiglaucoma medications needed for IOP control (⫾ standard deviation) was also reduced from a preoperative mean of 3.1 ⫾ 0.99 to 0.8 ⫾ 0.8 at 1 year (6 months if ⬍1 year follow-up, P ⬍ 0.001). Fourteen patients (58.3%) did not require any antiglaucoma medications at last follow-up. The IOP and medication summary are given in Table 2. The mean IOP and mean number of medications needed at specific time intervals after surgery are shown in Figure 2. Five patients had intraocular pressures of ⬍21 mmHg on their preoperative visit. Two of these patients had widely fluctuating intraocular pressures, which were sometimes ⬍21 mmHg. These two patients had stable intraocular pressures ⬍21 mmHg after the Baerveldt implant and were considered to have a successful outcome in the analysis. The other three patients had intraocular pressures between 18 and 19 mmHg on two to four antiglaucoma medications preoperatively. These three patients had IOP reductions of 28% to 42% on none to one antiglaucoma medication postoperatively and were also considered to have a successful outcome in the analysis.
Table 2. Intraocular Pressure and Medication Summary of Uveitic Glaucoma Patients Treated with Baerveldt Glaucoma Drainage Implants
IOP (mmHg) (mean ⫾ SD) Antiglaucoma medications (mean ⫾ SD) Zero One Two Three Four
Preoperative
1 Year Postoperative*
Last Follow-up†
30.5 ⫾ 8.96 3.1 ⫾ 0.99 1 0 4 9 10
13.0 ⫾ 4.6 0.8 ⫾ 0.8 12 6 6 0 0
13.3 ⫾ 5.1 0.8 ⫾ 1.1 14 4 5 0 1
* Six-months postoperative if ⬍ 1 year follow-up. † The mean postoperative follow-up was 20.8 months (range, 6 –52 months). IOP ⫽ intraocular pressure; SD ⫽ standard deviation.
2258
Ceballos et al 䡠 Baerveldt Implants for Uveitic Glaucoma Table 3. Postoperative Complications of Uveitic Glaucoma Patients Treated with Baerveldt Glaucoma Drainage Implants No. of Eyes 1
Description
Surgical Intervention
1*†
Limited suprachoroidal hemorrhage, hypotony maculopathy (transient), cystoid macular edema Kissing choroidal effusions, hypotony maculopathy, macular ischemia
1 1 1 1† 1* 1
Choroidal effusion (transient) Choroidal effusion (transient), hyphema Persistent low peripheral choroidal effusion Hypotony maculopathy, cystoid macular edema Cystoid macular edema Tube occluded with vitreous twice in eye with opaque cornea
1* 1
Failed corneal graft Failed corneal graft
None Choroidal effusions drained twice; tube permanently ligated None None None Religation of tube with absorbable suture None Tube repositioning anterior vitrectomy, penetrating keratoplasty, and tube repositioning Penetrating keratoplasty None
* Denotes patients who lost 2 or more lines of best-corrected visual acuity. Denotes the two surgical failures caused by chronic hypotony.
†
The preoperative best-corrected visual acuity ranged from 20/20 to hand motions. At last follow-up, 19 (79.2%) patients had a best-corrected visual acuity that was either unchanged (5 patients), improved (10 patients), or within 1 line (4 patients) of preoperative levels. Of the five patients who lost 2 or more lines of best-corrected visual acuity, the etiologies were as follows: macular scarring (one eye), macular ischemia related to kissing choroidal effusions (one eye), cystoid macular edema (one eye), amniotic membrane covering cornea at last follow-up (one eye), and failed corneal graft (one eye). There were no intraoperative complications in this series. Ten eyes had postoperative complications (Table 3). The most common were choroidal effusions (four eyes, 16.7%), hypotony maculopathy (three eyes, 12.5%), cystoid macular edema (three eyes, 12.5%), and failure of corneal grafts (two eyes, 8.3%). One patient in whom a corneal graft failed had undergone penetrating keratoplasty 3 months before placement of the Baerveldt implant. The corneal graft remained clear for 2 years and then slowly failed. The second patient in whom a corneal graft failed had undergone penetrating keratoplasty at the same time as placement of the Baerveldt implant. Her corneal graft failed 12 months after surgery, at which time she underwent a second penetrating keratoplasty. Most complications were transient and of no visual consequence. Those complications that resulted in loss of 2 or more lines of best-corrected visual acuity included kissing choroidal effusions with resulting macular ischemia (one patient, 6 lines of bestcorrected visual acuity lost), failure of corneal graft (one patient, 3 lines of best-corrected visual acuity lost), and cystoid macular edema (one patient, 2 lines of best-corrected visual acuity lost). There were no cases of tube– cornea touch, endophthalmitis, or plate extrusion. Seven of the 22 eyes (31.8%) in which the Baerveldt glaucoma drainage implant was considered a success required subsequent nonglaucoma-related ocular surgeries: excision of conjunctival inclusion cyst (one eye); phacoemulsification with posterior chamber intraocular lens implantation (one eye); superficial keratectomy with amnionic membrane transplantation (one eye), and penetrating keratoplasty alone or in combination with cataract extraction/IOL implantation, anterior vitrectomy, or tube repositioning (four eyes). All eyes that underwent post-Baerveldt penetrating keratoplasties had either a failed corneal graft (one eye) or decompensated corneas (three eyes) preoperatively. No eye that underwent incisional surgery after placement of the Baerveldt device subsequently lost control of IOP.
Discussion The best procedure for the surgical management of uveitic glaucoma refractory to medical therapy is still a subject of considerable debate. Although some studies report reasonable success rates using trabeculectomy with antimetabolites,1–5 in clinical situations involving conjunctival scarring, young patient age, a high likelihood of future ocular surgery, or a postoperative course likely to include protracted intraocular inflammation, placement of a glaucoma drainage device may be the preferred method for achieving long-term control of IOP. Our overall surgical success rate of 91.7% at 1 and 2 years with the Baerveldt glaucoma drainage implant is encouraging, especially in our highly varied patient population with multiple risk factors for failure of conventional filtering surgery: uveitis in all patients, aphakia or pseudophakia in 79.2%, black or Hispanic race in 66.7%, age ⬍60 years in 62.5%, and previously failed filtering surgery in 33.3% of patients. In addition, all the 10 patients who had failed trabeculectomy (8 patients) or previous tube shunt placement (2 patients) achieved successful IOP control with the Baerveldt device. Interestingly, none of the Baerveldt implants failed because of an inability to control elevated IOP postoperatively, and no surgeries failed after 6 months of follow-up, despite seven of the patients undergoing subsequent nonglaucoma-related incisional surgery. We also did not observe a trend with time for higher IOP or increased need for IOP-lowering medications. Although longer follow-up is needed to detect late failures, the lack of apparent attrition of success over time is encouraging, given that long-term studies of trabeculectomy with antimetabolites show that success decreases over time.9,10 Furthermore, the ability of the implant to function well despite subsequent surgical procedures is important, because these are frequently needed to deal with the complications of chronic intraocular inflammation or the steroids used for treatment. Both of the patients in whom the surgery failed did so early in the postoperative period and because of intractable
2259
Ophthalmology Volume 109, Number 12, December 2002 hypotony. These two patients were the youngest in the study (18 and 21 years old) and also had in common female gender, aphakia, and implant size (350 mm2). One patient developed kissing choroidal effusions that had to be drained twice, macular ischemia, and profound irreversible visual loss. Her tube necessitated permanent ligature. The second patient maintains an IOP in the 3.5 to 5.5 mmHg range (mean, 4.3 mmHg), with persistent inflammation, cystoid macular edema, and hypotony maculopathy. She has elected no treatment at this time. Unfortunately, the relatively small number of patients in our study and the common use of Baerveldt implants with larger surface areas (23 of 24 were 350 mm2 or larger) do not permit speculation as to whether a Baerveldt implant with a smaller surface area or a valved glaucoma drainage implant would have prevented the hypotony in these two patients. Da Mata et al6 also reported two young females (ages 17 and 28) who developed transient (one patient) and persistent (one patient) hypotony (IOP ⬍5) in their series of 19 patients who underwent placement of an Ahmed valve for the treatment of uveitic glaucoma. However, in a study by Valimaki et al7 of 27 eyes of 19 children who underwent placement of a Molteno implant (24 [89%] single plate and 3 [11%] double plate) for the treatment of secondary glaucoma associated with juvenile rheumatoid arthritis, no eyes developed persistent hypotony (IOP ⬍ 6 mmHg). Implantation of the Baerveldt glaucoma drainage device did not seem to deleteriously affect the uveitis: 12 patients who were using topical corticosteroids or nonsteroidal antiinflammatory agents preoperatively still required them at last follow-up, 5 patients who required topical steroids or topical nonsteroidal anti-inflammatory agents preoperatively did not require them at last follow-up, 5 patients who did not require topical steroids preoperatively required them at last follow-up, and a single patient needed the systemic steroids and immunosuppressants postoperatively in addition to her usual preoperative topical steroid regimen. Overall, the favorable results of this study indicate that implan-
2260
tation of the Baerveldt glaucoma drainage device is relatively safe and effective in eyes with uveitic glaucoma either as the initial surgical procedure for control of IOP or as an additional treatment modality for eyes that have failed previous glaucoma surgery.
References 1. Jampel HD, Jabs DA, Quigley HA. Trabeculectomy with 5-fluorouracil for adult inflammatory glaucoma. Am J Ophthalmol 1990;109:168 –73. 2. Prata JA Jr, Neves RA, Minckler DS, et al. Trabeculectomy with mitomycin C in glaucoma associated with uveitis. Ophthalmic Surg 1994;25:616 –20. 3. Wright MM, McGehee RF, Pederson JE. Intraoperative mitomycin-C for glaucoma associated with ocular inflammation. Ophthalmic Surg Lasers 1997;28:370 – 6. 4. Patitsas CJ, Rockwood EJ, Meisler DM, Lowder CY. Glaucoma filtering surgery with postoperative 5-fluorouracil in patients with intraocular inflammatory disease. Ophthalmology 1992;99:594 –9. 5. Towler HMA, Bates AK, Broadway DC, Lightman S. Primary trabeculectomy with 5-fluorouracil for glaucoma secondary to uveitis. Ocul Immunol Inflamm 1995;3:163–70. 6. Da Mata A, Burk SE, Netland PA, et al. Management of uveitic glaucoma with Ahmed glaucoma valve implantation. Ophthalmology 1999;106:2168 –72. 7. Valimaki J, Airaksinen JA, Tuulonen A. Molteno implantation for secondary glaucoma in juvenile rheumatoid arthritis. Arch Ophthalmol 1997;115:1253– 6. 8. Molteno ACB, Sayawat N, Herbison P. Otago glaucoma surgery outcome study: long-term results of uveitis with secondary glaucoma drained by Molteno implants. Ophthalmology 2001;108:605–13. 9. Araie M, Shoji N, Shirato S, Nahano Y. Postoperative subconjunctival 5-fluorouracil injections and success probability of trabeculectomy in Japanese: results of 5-year follow-up. Jpn J Ophthalmol 1992;36:158 – 68. 10. Three-year follow-up of the Fluorouracil Filtering Surgery Study. Am J Ophthalmol 1993;115:82–92.