Ninety-four patients with congenital aniridia were identified, of which 46 (49%) were male. Median age at presentation was 19.0 years (mean 23.7 ± 21.6 years, range 0–84). Self-identified racial ethnicity was Caucasian (87%), Black (9%), Asian (3%), and Hispanic (1%). Fifty-one patients (54%) had positive family history (familial) while 28 (30%) were documented to have no family history (sporadic). The remaining 15 patients (16%) did not have family history documented in the medical record. There were no differences in sex (p = 0.85), race (p = 0.85) or follow up length (p = 0.235) amongst aniridic patients classified as familial, sporadic, and unknown family history.
Twelve patients were members of a previously described 4-generation family, in which genetic testing showed a splice-site mutation (c.565TC > T) in PAX6 (NM_000280.4).[18] Seven additional patients were related to this family by marriage (Fig. 1). Other familial relationships within our cohort included: father with 3 children, mother with 2 children, father and son, brother and sister twin siblings, and 2 sister pairs. In addition to the aforementioned splice-site mutation, 1 patient with familial aniridia had 11p14.1-p13 deletion (951Kb loss) and 1 patient had 8p23.2 duplication of unknown significance (Supplemental Table 1). Within the sporadic group, 4 patients had a confirmed mutation involving PAX6 or the neighboring ELP gene. Five patients with WAGR syndrome had gene deletions or duplications involving 11p14. Interestingly, 1 patient who clinically had WAGR syndrome had a frameshift mutation reportedly affecting PAX6, but not adjacent genes.
At median follow-up of 4.0 years (mean 8.7 ± 12.0 years, range 0.1–53.8), 112 eyes (60%) of 62 patients (66%) underwent at least 1 surgery. Of these patients, 12 had unilateral surgery while 50 had bilateral surgeries. The average number of surgeries was 1.74 ± 2.30/eye (median 1, range 0–14). At presentation, the crystalline lens was present in 125 eyes (66%) of 66 patients (70%), of which 79 of 125 (63%) had cataract (Fig. 2A). Forty-nine eyes (26%) of 31 patients(33%) had previously undergone lensectomy, with 30 eyes (59%) having intraocular lens (IOL) implantation and 19 eyes (37%) remaining aphakic. Three eyes (2%) of 3 patients (3%) had enucleations prior to presentation and 9 eyes (5%) of 7 patients (7%) had unknown lens status due to corneal opacification. At final follow-up, 91 eyes (48%) of 52 patients (55%) remained phakic, although 61 eyes had a cataract and 3 eyes showed partial lens dislocation. Thus, an additional 36 eyes underwent lensectomy such that at final follow-up 85 eyes (45%) of 51 patients (54%) were pseudophakic and 30 eyes (16%) of 20 patients (21%) were aphakic. Only 1 eye had capsular tension ring (CTR) placement during lensectomy. Four eyes of 3 patients initially had IOL placement, which was later removed due to dislocation. Two additional eyes were enucleated during follow-up. Seven eyes (4%) of 6 patients (6%) had corneal opacification that precluded lens status at final follow-up.
At presentation, 75 eyes (40%) of 39 patients (41%) had glaucoma, of which 17 eyes (23%) of 12 patients (31%) had undergone prior IOP-lowering surgery (Fig. 2B). At final follow-up, 98 eyes (52%) of 50 patients (53%) had glaucoma, of which 49 eyes of 28 patients required IOP-lowering surgery with an average of 1.9 ± 1.6 surgeries/eye (range 1–10, median 1). Due to the extensive length of follow-up and conversion from paper to electronic records, accurate survival times were unable to be calculated for 54 of 94 glaucoma surgeries performed. Survival times of less than half of the surgeries would not yield accurate results, however, success rates at final follow-up were determined (Table 1). Success at final follow-up of Baerveldt and Ahmed GDDs were 74% and 63%, respectively. This increased to 87% and 88% if failure criteria did not include tube revision. Success of trabeculectomy with anti-fibrotics was 24%, which increased to 33% if bleb revision surgeries were excluded. Angle surgery showed 33% success. Success of cycloablation was 33% which increased to 50% if multiple cycloablations were excluded. There was no significant difference in length of follow-up between glaucoma surgeries (p > 0.2).
Table 1
Glaucoma Surgery (n) | Previous Glaucoma Surgeries 0 1 >2 surgeries surgery surgeries | Success at Final Follow-up | Success Exclusion | Success with Exclusion at Final Follow-up | Length of Follow-up (yrs) |
Glaucoma Drainage Device (35) Baerveldt (23) Ahmed (8) Unspecified (4) | 23 eyes 7 eyes 5 eyes 16 eyes 4 eyes 3 eyes 3 eyes 3 eyes 2 eyes 4 eyes | 71% (25 of 35 Eyes) 74% (17 of 23 Eyes) 63% (5 of 8 Eyes) 75% (3 of 4 Eyes) | Tube Revision | 89% (31 of 35 Eyes) 87% (20 of 23 Eyes) 88% (7 of 8 Eyes) 100% (4 of 4 Eyes) | 14.2 ± 15.4 13.2 ± 14.3 17.5 ± 20.1 2.7 ± 3.6 |
Trabeculectomy with Anti-Fibrotics (21) | 14 eyes 6 eyes 1 eye | 24% (5 of 21 Eyes) | Bleb Revision | 33% (7 of 14 Eyes) | 21.7 ± 14.1 |
Angle Surgery (9) Trabeculotomy (5) Goniotomy (4) | 8 eyes 1 eye 4 eyes 1 eye 4 eyes | 33% (3 of 9 Eyes) 40% (2 of 5 Eyes) 25% (1 of 4 Eyes) | Repeat Angle Surgery | 33% (3 of 9 Eyes) 40% (2 of 5 Eyes) 25% (1 of 4 Eyes) | 16.3 ± 17.8 26.6 ± 15.3 0.9 ± 1.3 |
Cycloablation (12) Contact Transcleral (5) Endoscopic (6) Cryo (1) | 3 eyes 4 eyes 5 eyes 1 eye 1 eye 3 eyes 2 eyes 3 eyes 1 eye 1 eye | 33% (4 of 12 Eyes) 40% (2 of 5 Eyes) 17% (1 of 6 Eyes) 100% (1 of 1 Eye) | Multiple Cyclo- Ablations | 50% (6 of 12 Eyes) 60% (3 of 5 Eyes) 33% (2 of 6 Eyes) 100% (1 of 1 Eye) | 23.4 ± 12.0 19.8 ± 7.4 20.5 ± 8.6 43.4 |
At presentation, 108 eyes(57%) of 56 patients(60%) had keratopathy, including 1 eye which had already undergone Boston keratoprosthesis (KPro) implantation (Fig. 2C). At final follow-up, 123 eyes (65%) of 63 patients (67%) had keratopathy, of which 44 eyes (36%) of 28 patients (44%) underwent at least 1 corneal procedure. The average number of corneal surgeries was 1.7 ± 1.1/eye (range 1–6, median 1). Corneal surgeries included penetrating keratoplasty (PKP, 19 eyes), Kpro (18 eyes), limbal stem cell transplantation (LSCT, 12 eyes), lamellar keratoplasty (LK, 4 eyes), and Descemet stripping endothelial keratoplasty (DSEK, 3 eyes). Kaplan-Meier analysis showed that PKP (Fig. 3A) had 1, 10, and 25-year survival rates of 90% with 95% CI [71, 96], 43% with 95% CI [23, 62], and 20% with 95% CI [6, 40], respectively. Of the 18 eyes which underwent KPro, 11 had not undergone prior corneal surgery while 6 had a history of failed PKP and 1 had previous LK. KPro survival rates were 95% with 95% CI [70, 99] at 1-year, 83% with 95% CI [55, 94] at 5-years, and 64% with 95% CI [32, 84] at 10-years. DSEK (Fig. 3B) 1-year survival rate was 75% with 95% CI [13, 96] and decreased to 25% with 95% CI [1, 66] at 2 years. LK had 5-year, 10-year, and 15-year survival rates of 88% with 95% CI [40, 99], 63% with 95% CI [23, 87], and 31% with 95% CI [6, 64], respectively. LSCT showed a 1-year survival rate of 75% with 95% CI [13, 96] and 5-year survival rate of 25% with 95% CI [1, 66]. Log-rank (Mantel Cox) analysis showed a significant difference between survival curves of the aforementioned corneal surgeries (p = 0.0006).
Additional surgeries included pars plana vitrectomy (PPV) in 20 eyes (11%) and artificial iris implantation in 4 eyes (2%). In 13 eyes, PPV was performed in conjunction with GDD implantation (7 eyes) and/or KPro placement (6 eyes). Two additional eyes had PPV combined with trabeculectomy or PKP. Five eyes underwent PPV to repair retinal detachments that were due to prior surgery (3 eyes), trauma (1 eye) or an optic nerve pit (1 eye). One eye, in addition to PPV, also had placement of scleral buckle for a KPro-related retinal detachment.
LogMAR BCVA at presentation (1.46 ± 0.46, median 1.0) was not significantly different (p = 0.51) than at final follow-up (1.32 ± 0.76, median 1.10). At final follow-up, 57 eyes (30%) had BCVA better than 20/200 with 11 (6%) having BCVA of 20/50 or better (Fig. 4A). In contrast, 107 eyes (57%) had BCVA worse than 20/200 with 16 eyes (9%) with no light perception. Twenty-four eyes (13%) were too young to cooperate with optotype testing at final follow-up. Analysis of the better seeing eye in each patient (Fig. 4B) showed similar results with 38% with BCVA better than 20/200. However, 50% met the legal definition of blindness in the better seeing eye at final follow-up. Optotype visual acuity testing was completed at both initial and final visits in 133 eyes (71%) of 67 patients (71%). Ninety-three eyes (50%) had stable (< 0.1 change in LogMar VA) or improved BCVA (Fig. 4C) while 40 eyes (21%) showed worse BCVA at final follow-up. Likewise, BCVA of the better seeing eye (Fig. 4D) improved or remained stable in 52% of patients, but decreased in 19% of patients.
Univariate analysis of all 188 eyes showed that final BCVA was associated with BCVA at initial presentation (p < 0.001), age at initial presentation (p < 0.0001), glaucoma diagnosis (p < 0.001), cataract at presentation (p = 0.037), and presence of keratopathy at presentation (p < 0.0001). While there was no difference in final BCVA between phakic and pseudophakic eyes, aphakic eyes showed worse BCVA compared to pseudophakic (p < 0.005) and phakic (p < 0.005) eyes. Furthermore, eyes which underwent PKP or KPro implantation were associated with worse BCVA (p < 0.0001 and p < 0.05, respectively). There was no relationship between final BCVA and LSCT, glaucoma surgery, retina surgery, or iris prosthetic implantation. Neither sex nor race were associated with final BCVA. No variables modified sex or race in order to predict worse final BCVA.