To assess the effect of Ahmed glaucoma valve implants on corneal endothelial cell density (ECD) in children with uveitic glaucoma.
setting : Institutional. patient population : Eighty eyes from 42 patients diagnosed with uveitis before the age of 16. Twenty-eight eyes had an Ahmed glaucoma valve implant because of secondary glaucoma. Fifty-two eyes without an implant served as controls. intervention or observation procedure(s) : Corneal ECD was examined cross-sectionally using a noncontact specular microscope. Univariate and multivariate generalized estimating equations analyses with correction for paired eyes were performed. main outcome measure(s) : Correlation of ECD with the presence of an Ahmed glaucoma valve implant and with the time following implantation.
ECD was significantly lower in the Ahmed glaucoma valve group than in controls (2359 and 3088 cells/mm 2 , respectively; P < .001) following an average of 3.5 years after Ahmed glaucoma valve implantation. Presence of an Ahmed glaucoma valve implant, previous intraocular surgery, age, duration of uveitis, and history of corneal touch by the implant tube were all significantly associated with decreased ECD. Following a multivariate analysis, presence of an Ahmed glaucoma valve implant (B = −340; adjusted P < .011) and older age (B = −58; adjusted P = .005) remained independently associated with decreased ECD. Within the implant group, the age-adjusted time interval following Ahmed glaucoma valve implantation was highly correlated with decreased ECD (B = −558, P < .001).
Ahmed glaucoma valve implants in children with uveitic glaucoma are independently associated with decreased ECD, and this effect is associated with the time interval following Ahmed glaucoma valve implantation.
Uveitis is an important cause of ocular morbidity in children and accounts for 5%-20% of blindness cases in the United States and Europe. Glaucoma can arise secondary to uveitis or the corticosteroids used to treat uveitis. Uveitic glaucoma is a common complication that occurs within 5 years among 12%-46% of children with all types of uveitis. Legal blindness can be attributed to secondary glaucoma in 15% of children with all types of uveitis, and this prevalence rises to 50% among children with uveitis-associated with juvenile idiopathic arthritis (JIA).
Children with uveitic glaucoma carry a high risk for trabeculectomy failure. The relatively young age, ocular inflammation, and frequent need for cataract surgery among these patients all predispose the patient to excessive fibrosis, which is associated with a less favorable outcome following trabeculectomy. Glaucoma implant valve devices such as an Ahmed glaucoma valve or Baerveldt implant have been proven effective for lowering intraocular pressure in refractory glaucoma. However, their long-term effects on the corneal endothelium in inflammatory glaucoma are unknown.
Corneal endothelial cells do not proliferate in humans. Endothelial cell density (ECD) is highest after birth (with a mean cell density of 4000-6000 cells/mm 2 ), after which it declines rapidly in the first 2 years of life. Known factors for endothelial cell damage include cataract extraction, glaucoma surgery, and corneal transplantation. Among the available topical drugs, only dorzalomide (a carbonic anhydrase inhibitor) seems to have a possible negative effect on the corneal endothelium.
Two recent studies prospectively evaluated the effect of Ahmed glaucoma valve implants on ECD following 1 and 2 years of follow-up and suggested a progressive corneal endothelial cell loss in the eyes that received an implant. However, neither of these studies included children, and only a portion of the study populations had glaucoma secondary to uveitis. Therefore, the purpose of the present study was to investigate the effect of an Ahmed glaucoma valve implant on corneal endothelial cells in children with glaucoma secondary to uveitis.
In this cross-sectional observational case study, we included all patients who were diagnosed with uveitis before the age of 16 years and who presented with glaucoma secondary to uveitis and received an Ahmed glaucoma valve implant at the Department of Ophthalmology at the University Medical Center, Utrecht, The Netherlands, from March 10, 2005 through December 5, 2011. Children with uveitis who were diagnosed before the age of 16 and who did not receive an Ahmed glaucoma valve implant served as the control group. The unaffected eyes of unilateral uveitis patients and eyes with phthisis bulbi secondary to longstanding uveitis were excluded. This study was approved by the University Medical Center Utrecht’s Institutional Review Board and was performed in accordance with the 1990 Declaration of Helsinki and subsequent amendments. Informed consent for the evaluation of the endothelial cell counts and for using their data in this study was obtained from the patients’ parents and from children who were older than 12 years of age.
Corneal endothelial cells were examined by experienced physicians using a noncontact specular microscope (Topcon Specular Microscope model SP-2000P; TOPCON corporation, Tokyo, Japan). For each patient, corneal endothelial cell density was assessed once in the central region. A diagnosis of uveitis was based on the criteria of the Standardization of Uveitis Nomenclature (SUN) Working Group. A diagnosis of JIA was made in accordance with the criteria established by the International League against Rheumatism (ILAR). Each preliminary presumptive diagnosis of JIA was confirmed by a pediatric rheumatologist. Diagnoses of underlying systemic diseases associated with uveitis were established in accordance with current diagnostic criteria.
The following clinical data were collected from each patient’s medical records: sex, date of the onset of uveitis, anatomic type and etiology of uveitis, previous cataract extraction and/or trabeculectomy, topical administration of carbonic anhydrase inhibitors (because of their potential effects on endothelial cell function), preoperative administration and application duration of wound-healing modulators such as 5-fluorouracil and mitomycin C, and date and eventual complications of the first Ahmed glaucoma valve implantation, such as implant failure or tube-related complications. Previous cataract extraction and/or trabeculectomy were considered to be a previous intraocular surgery.
All surgeries were performed by specialized physicians under general anesthesia. The Ahmed glaucoma valve devices (model S2 [surface area of 184 mm 2 ] or model FP8 [surface area of 96 mm 2 ], New World Medical Inc, Rancho Cucamonga, California, USA) were implanted using a standard technique, and each valve was placed in the superior temporal or superior nasal area. The indication for an FP8 implant was a previous period of hypotony. Each patient was pretreated with systemic corticosteroids (1 mg/kg) beginning 2 days prior to surgery and received periocular betamethasone during surgery. Postoperatively, the systemic corticosteroids were slowly tapered off based on the activity of the intraocular inflammation.
Statistical analysis was performed “by eye” using SPSS version 20.0 for Windows (SPSS Inc, Chicago, Illinois, USA). Generalized estimating equations with statistical correction to test for correlations between paired eyes were used for univariate and multivariate analyses. Correlations between variables with correction for paired eyes were also analyzed using generalized estimating equations. Differences with P ≤ .05 were considered to be statistically significant. Variables that were found to be statistically significant in the univariate analysis were subsequently included in the multivariate analysis.
Eighty eyes from 42 patients were included in this study. The characteristics of the study subjects are summarized in Table 1 . The Ahmed glaucoma valve group contained 28 eyes from 21 patients with an S2 implant (n = 26 eyes) or an FP8 implant (n = 2 eyes). Seven and 14 patients had bilateral and unilateral Ahmed glaucoma valve implants, respectively. The control group consisted of 52 eyes without an Ahmed glaucoma valve and included eyes with bilateral uveitis (n = 36), eyes with unilateral uveitis (n = 3), and the unaffected eyes of patients with a unilateral Ahmed glaucoma valve device (n = 13; 1 unaffected eye of a patient with a unilateral Ahmed glaucoma valve had phthisis and was therefore excluded from the analysis) ( Table 1 ). Both the mean age and the average number of previous intraocular surgeries in the Ahmed glaucoma valve group were significantly higher than in the control group ( Table 1 ). There were no statistical differences between the 2 groups with respect to the other demographic characteristics ( Table 1 ). The only subsequent intraocular surgeries in the eyes in the Ahmed glaucoma valve group were revisions of the implants in 6 eyes for the following reasons: tube revision because of corneal touch (n = 2), failure of the implant (n = 2), hypotonia (n = 1), and retraction of the tube (n = 1); the mean time until revision was 2.4 years.
|Characteristics a||Ahmed Group (N of Eyes = 28)||Control Group (N of Eyes = 52)||P Value b|
|Number of patients (N = 42)||21||34 c||NA|
|Mean ± SD age (y)||16.7 ± 5.0||12.7 ± 4.6||.011|
|Sex, n (%)||.220|
|Male||11 (39%)||22 (42%)|
|Female||17 (61%)||30 (58%)|
|Anatomic type, n (%)||.169|
|Anterior||24 (86%)||38 (73%)|
|Intermediate||3 (11%)||11 (21%)|
|Posterior||0 (0%)||1 (2%)|
|Panuveitis||1 (4%)||2 (4%)|
|Etiology, n (%)||.292|
|JIA||25 (89%)||40 (77%)|
|Sarcoidosis||1 (4%)||3 (6%)|
|Toxocara||0 (NA)||1 (2%)|
|Unknown||2 (7%)||8 (16%)|
|Mean ± SD age at uveitis diagnosis (y)||4.7 ± 2.1||5.3 ± 2.3||.882|
|Mean ± SD duration of uveitis (y)||12.0 ± 5.4||7.4 ± 4.8||.888|
|Previous intraocular surgery, n (%) d||<.001|
|None||0 (NA)||26 (50%)|
|Cataract extraction e||26 (93%)||21 (23%)|
|Trabeculectomy with wound healing modulators||14 (50%)||14 (27%)|
|Administration of topical carbonic anhydrase inhibitor, n (%)||10 (36%)||23 (44%)||.474|
|Mean ± SD age at implantation of Ahmed glaucoma valve (y)||13.2 ± 4.4||NA||NA|
|Mean ± SD time following Ahmed glaucoma valve implantation (y)||3.5 ± 2.2||NA||NA|
c Including 13 patients with unilateral Ahmed glaucoma valve implant, whose unaffected eyes were included in the control group. One unaffected eye of a patient with a unilateral Ahmed glaucoma valve had phthisis and was therefore excluded from the analysis.
e Lens status in the Ahmed glaucoma valve group and the control groups: phakic (n = 2 and n = 31, respectively), pseudophakic (n = 21 and n = 20, respectively), and aphakic (n = 5 and n = 1, respectively).
The mean ± SD postimplant (an average of 3.5 years following implantation) ECD count in the Ahmed glaucoma valve group was 2359 ± 831 cells/mm 2 (range: 864-3810 cells/mm 2 ). In contrast, the ECD count in the control group was 3088 ± 442 cells/mm 2 (range: 2085-4195 cells/mm 2 ), which was significantly higher than in the Ahmed glaucoma valve group ( P < .001) ( Figure 1 ).
A univariate analysis revealed that a lower ECD was significantly associated with the following factors: the presence of an Ahmed glaucoma valve implant, previous intraocular surgery, age, duration of uveitis, and corneal touch ( Table 2 ). In contrast, revision of the Ahmed glaucoma valve implant, perioperative application of wound-healing modulators, and the administration of topical carbonic anhydrase inhibitors were not associated with a lower ECD count ( Table 2 ).
|B||95% CI||Crude P Value||B||95% CI||Adjusted P Value|
|Presence of an Ahmed glaucoma valve implant||−619||−939, −300||<.001||−340||−603, −77||.011|
|Age||−77||−121, −33||.001||−58||−99, −18||.005|
|Previous intraocular surgery b||−480||−784, −76||.002||−140||−329, 49||.147|
|Corneal touch||−1442||−2640; −243||.018||−1026||−2096, 45||.061|
|Duration of uveitis c||−57||−94, −20||.003||Not applicable c|
|Administration of wound-healing modulators||−241||−496, 12||.062||Not applicable|
|Revision of Ahmed glaucoma valve implant||−245||−943, 451||.490||Not applicable|
|Administration of carbonic anhydrase inhibitors||−37||−286, −212||.771||Not applicable|