To evaluate the efficacy of a second glaucoma implant in eyes with prior glaucoma implant surgery and inadequate intraocular pressure (IOP) control.
Retrospective observational cohort study.
Patients undergoing a second glaucoma implant surgery from 1996 to 2008 were included. Outcome measures included visual acuity, IOP, glaucoma medication use, and complications. Success was defined as IOP < 21 mm Hg (criterion 1) and IOP < 17 mm Hg (criterion 2), with at least 25% reduction in IOP and no prolonged hypotony.
Forty-three eyes (43 patients) had a mean follow-up of 32.6 ± 21.6 months. Life-table analysis demonstrated success rates of 93%, 89%, and 83% using criterion 1 and 83%, 75%, and 75% using criterion 2 at 1, 2, and 3 years, respectively. At last follow-up, mean IOP (13.6 ± 4.6 vs 24.7 ± 7.5 mm Hg; P < .001) and mean number of medications (1.4 ± 1.2 vs 3.9 ± 1.2; P < .001) were lower following the second implant. There was no difference in preoperative and most recent logarithm of the minimal angle of resolution (logMAR) visual acuities (0.86 ± 0.13 vs 1.1 ± 0.13; P = .07). The most frequently used second implants were similar in percentage IOP reduction (Baerveldt implant, 45 ± 19%; Ahmed valve, 40 ± 18%; P = .4).
A second glaucoma implant may effectively lower IOP in eyes with refractory glaucoma.
The most common current indication for glaucoma drainage device (GDD) implantation is uncontrolled glaucoma with intraocular pressure (IOP) refractory to medical therapy, laser treatment, and filtering surgery. Successful outcome following GDD implantation has been reported in 50% to 90% of patients depending on the severity of glaucoma, length of follow-up, and definition of success.
A glaucoma implant may fail to adequately control IOP for some patients. Failure to achieve the target IOP is most often attributable to excessive fibrosis around the reservoir. Shunt revision surgery has been attempted for these eyes with limited long-term success rates. However, in many cases with inadequate IOP control following primary GDD implantation, surgical options may be limited to cyclophotocoagulation or placement of a second aqueous shunt.
Although there are several reports of a favorable outcome after cyclodestructive procedures in refractory glaucomas with good visual potential, many surgeons are hesitant to destroy the ciliary body because of the risks of uveitis, vision loss, hypotony, and phthisis. Complication rates have been reported to be relatively lower with endocyclophotocoagulation (ECP) compared with transscleral cyclophotocogulation (TCP). However, large studies with long-term follow-up are lacking. Alternatively, implantation of a second GDD may be performed. Data regarding the efficacy and potential complications associated with the implantation of an additional GDD is limited to three small retrospective studies published between 2000 and 2002. The purpose of this report is to assess the effectiveness and complications of sequential GDD implantation in consecutive patients with refractory glaucoma undergoing this procedure.
We reviewed the surgical records of all patients undergoing GDD implantation at the New York Eye and Ear Infirmary between January 1, 1996 and January 1, 2008 to identify patients with sequential tube implants. These patients had undergone a second tube implant for inadequate IOP control despite the use of maximal tolerated medical therapy. Inclusion criteria included both implant surgeries having been performed by a glaucoma specialist at the New York Eye and Ear Infirmary and at least 12 months of follow-up after the second surgery. One eye was randomly selected for patients with bilateral sequential tube implants.
Demographic, clinical, and operative details were extracted from the medical record. Details of IOP control, glaucoma medications, surgical complications, and visual acuity (VA) were collected at baseline assessment, before the second implant, and at each postoperative visit. The interval from baseline assessment to the first aqueous shunt, the first to second aqueous shunt, and the second shunt to most recent follow-up were determined.
The primary outcome measure was IOP control. Secondary outcome measures included number of glaucoma medications, VA, and surgical complications. Success was defined as IOP < 21 mm Hg with at least 25% reduction in IOP on equal or less number of antiglaucoma drops and no prolonged hypotony (IOP < 4 mm Hg persisting for more than 8 weeks and/or associated hypotony related complications). A second criterion for success was defined as IOP < 17 mm Hg with at least 25% reduction in IOP on equal or less number of antiglaucoma drops and no prolonged hypotony.
Snellen VA was converted to logarithm of the minimal angle of resolution (logMAR) equivalent for purpose of analysis. IOPs were reported both as absolute values and as percentage reduction from baseline. The mean of the last three IOP measurements taken before GDD surgery was used to report preoperative IOP. The means of IOP measurements between 3 and 6 months postoperatively and between 6 and 12 months postoperatively were used to report IOP at 6 months and 12 months, respectively. IOPs for subsequent postoperative years were reported by averaging all IOP measurements during that year. Measurements in the first 3 postoperative months were not used to calculate mean IOP for any time point.
Statistical analysis was performed using paired two-sided t test or analysis of variance for continuous variables and χ 2 /Fisher exact test for categorical variables. Life-table analysis was performed using Kaplan-Meier estimates. A P value of .05 or less was considered statistically significant. To compare groups based on implant types, our sample size of 43 eyes could detect an 8% difference in mean percentage IOP reduction and a 3.75 mm Hg difference in mean IOP with different implant types with a two-sided level of significance of .05 and a power of 0.80.
Fifty-five eyes of 52 patients underwent sequential tube implantation between January 1, 1996 and January 1, 2008. Of these, 43 eyes met the inclusion criteria and were selected for analysis. Five excluded eyes had less than 12 months follow-up and 4 excluded eyes had the initial glaucoma implant surgery performed elsewhere. Table 1 shows the demographic and clinical profiles of the study population. Mean number of incisional intraocular surgeries before any tube implants was 2.6 ± 1.2. Previous surgeries included failed trabeculectomy in 41 (95.3%) eyes, cataract surgery in 39 (90.7%) eyes, penetrating keratoplasy (PK) in 4 (9.3%) eyes, pars plana vitrectomy in 7 (16.3%) eyes, and goniotomy in 2 (4.7%) eyes. Repeat trabeculectomy or trabeculectomy revision had been performed in 19 (44.2%) eyes. Cataract surgery had been performed by a phacoemulsification procedure through clear corneal incision in 33 (84.6%) eyes, an extracapsular cataract extraction in 3 (7.7%), eyes and a pars plana lensectomy in 3 (7.7%) eyes. Initial GDD implantation included Baerveldt implant (Advanced Medical Optics Inc, Santa Ana, California, USA) in 22 (51.2%) eyes, Ahmed valve (New World Medical Inc, Rancho Cucamonga, California, USA) in 14 (32.6%) eyes, Molteno implant (Molteno Ophthalmics Ltd, Dunedin, New Zealand) in 5 (11.6%) eyes, Krupin valve (Eagle Vision Inc, Memphis, Tennessee, USA) in one eye, and Schocket implant in one eye. The mean duration between the first and second GDD surgery was 47.9 ± 27.7 months. Second implants included Baerveldt implant in 24 (55.8%) eyes, Ahmed valve in 17 (39.5%) eyes, and Molteno implant in 2 (4.7%) eyes. The sequential GDD implantation was a second glaucoma surgery in 2 (4.7%) eyes, third glaucoma surgery in 21(48.8%) eyes, fourth glaucoma surgery in 19 (44.2%) eyes, and fifth glaucoma surgery in 1 (2.3%) eye. The mean follow-up following second tube implants was 32.6 ± 21.6 months (range, 12 to 76 months). The details of shunt types, plate location, and tube placement are given in Table 2 .
|Mean age in years (range)||52.8 ± 22.9 (5 to 87)|
|Secondary OAG||6 (14.0%)|
|Uveitic glaucoma||4 (9.3%)|
|Congenital glaucoma||4 (9.3%)|
|Angle recession||3 (7.0%)|
|Mixed-mechanism glaucoma||4 (9.3%)|
|ICE syndrome||1 (2.6%)|
|First Implant||Second Implant|
|Type of implants|
|Ahmed||14 (32.6%)||17 (39.5%)|
|Baerveldt||22 (51.2%)||24 (55.8%)|
|Molteno a||5 (11.6%)||2 (4.7%)|
|Location of tube|
|Anterior chamber||32 (74.4%)||22 (51.2%)|
|Sulcus/vitreous cavity||11 (25.6%)||21 (48.8%)|
|Location of plate|
|Supero-temporal||37 (86.0%)||2 (4.7%)|
|Supero-nasal||3 (7.0%)||17 (39.5%)|
|Infero-temporal||2 (4.7%)||8 (18.6%)|
|Infero-nasal||1 (2.3%)||16 (37.2%)|
Baseline IOP before any device implantation was 31.2 ± 10.0 mm Hg (range, 16 to 54 mm Hg). Mean number of antiglaucoma drops being used were 3.6 ± 1.4 (range, 0 to 6). Twenty (46.5%) patients were using systemic carbonic anhydrase inhibitors. Mean IOP before the second GDD implantation was 24.7 ± 7.5 mm Hg (range, 11 to 50 mm Hg). Mean number of antiglaucoma drops being used before the second implant was 3.9 ± 1.2 (range, 0 to 6). Twenty-five (58.1%) patients were using systemic carbonic anhydrase inhibitors. Mean IOP and mean number of antiglaucoma drops at last follow-up following second implant were 13.6 ± 4.6 mm Hg (range, 6 to 25 mm Hg) and 1.4 ± 1.2 (range, 0 to 4), respectively. The mean percent reduction in IOP following the second implant was 44.1 ± 18.1% (range, 4.7% to 82.2%). Four (9.3%) patients were on systemic carbonic anhydrase inhibitors at the last follow-up. Mean IOPs and mean number of antiglaucoma medications were significantly less at all intervals following second GDD implantation ( Table 3 ).
|Duration||IOP (mm Hg)||Glaucoma Drops (n)|
|Mean||Range||P value||Mean||Range||P value|
|Preoperative||24.7 ± 7.5||11.0 to 50.0||3.9 ± 1.2||0 to 6.0|
|6 m||15.1 ± 4.4||5.3 to 24.3||<.001||1.4 ± 1.1||0 to 4.0||<.001|
|12 m||13.7 ± 5.0||5.0 to 23.0||<.001||1.5 ± 1.1||0 to 4.0||<.001|
|24 m||13.2 ± 4.3||5.0 to 20.0||<.001||1.8 ± 1.2||0 to 4.0||<.001|
|36 m||12.2 ± 4.5||5.2 to 18.0||<.001||1.4 ± 1.1||0 to 4.0||<.001|
|48 m||13.2 ± 5.2||8.7 to 17.5||.001||1.5 ± 1.2||0 to 4.0||<.001|
|60 m||13.4 ± 2.8||8.7 to 17.5||.01||1.5 ± 1.1||0 to 3.0||<.01|
|72 m||11.7 ± 4.0||7.0 to 15.8||.03||1.8 ± 1.0||1.0 to 3.0||.07|
Kaplan-Meier life-table survival analysis was used to calculate success probabilities ( Figure 1 ). Using outcome criterion 1, the probability of successful outcome was 92.9%, (95% confidence interval [CI], 99.8% to 84.8%; n = 43), 88.8% (95% CI, 99.6% to 75.6%; n = 39), and 83.2% (95% CI, 99.1% to 64.4%; n = 23) at 1, 2, and 3 years, respectively, following the second GDD surgery. Using outcome criterion 2, the probability of successful outcome was 83.3% (95% CI, 95.7% to 71.0%; n = 43), 75.3% (95% CI, 94.8% to 59.4%; n = 35), and 75.3% (95% CI, 98.6% to 54.4%; n = 20) at the same intervals.
The most common sequential implant combinations in our series were Baerveldt-Ahmed in 11 (25.6%) eyes, Baerveldt-Baerveldt in 11 (25.6%) eyes, and Ahmed-Baerveldt in 11 (25.6%) eyes. Other implants combinations were Ahmed-Ahmed in 2 eyes, Ahmed-Molteno in 1 eye, Molteno-Baerveldt in 2 eyes, Molteno-Ahmed in 3 eyes, Krupin-Ahmed in 1 eye, and Schocket-Molteno in 1 eye. The most frequently used GDD combinations were similar in percentage IOP reduction from IOP before any device implantation (Baerveldt-Ahmed, 55%±14%; Ahmed-Baerveldt, 55%±18%; Baerveldt-Baerveldt, 53%±19%; P = .9) ( Figure 2 ). The most frequently used second implants were similar in percentage IOP reduction from preoperative IOP before second implant surgery (Baerveldt, 45%±19%; Ahmed, 40%±18%; P = .4) ( Figure 2 ). The combined plate surface area of both implants ranged from 325 mm 2 to 700 mm 2 . No correlation was seen between the total combined implant surface area and percentage reduction in IOP ( Figure 3 ).