At present, there is no generally accepted or recommended universal surgical approach for the management of juvenile uveitic secondary glaucoma. Surgical treatment modalities include trabeculectomy with MMC, goniotomy, trabeculodialysis implantation of drainage devices and cyclodestructive procedures. The success rates reported for these techniques at the end of profoundly diverse follow-up times vary between 60% and 90% [5, 11–14]. One study focused on JRA patients, which is a high risk glaucoma group comparable to our patients. In this group, which is known for its high failure rate after surgery, conventional filtering surgery was able to control IOP in 57%, while trabeculectomy with MMC controlled IOP in the remaining four cases [5]. Cyclodestructive procedures, such as transscleral diode laser cyclophotocoagulation, proved ineffective as a primary surgical approach in JIA-associated uveitis and secondary glaucoma as the success rate was only 32% after 1 year [15].
The value of non-penetrating glaucoma surgery for the management of uveitic glaucoma has only been described in few reports on adult patients. Overall success rates (complete and qualified success) range from 87% to 100%. A favourable aspect of this technique is that complications (in these cases, lens opacities, reversible hypotony, hyphema and bleb encapsulation) occur infrequently [16–18]. We used a slightly modified standard deep sclerectomy to obtain a conventional transscleral deep filtration and combined it with circumscribed goniotomies to increase the filtration. In order to minimize the risk of vitreous prolapse under the flap in aphakic eyes, the incisions on the inner wall of Schlemm's canal were placed at each side. We speculated that this might prevent the need for subsequent goniopuncture, a procedure commonly required after deep sclerectomy [19]. Until now, no studies have been available presenting data on non-penetrating glaucoma surgery and standard trabeculectomy in uveitis patients, especially in children. Souissi and co-workers have published their data on deep sclerectomy and on trabeculectomy in two separate papers in adults. Their groups consisted of 8 patients in the paper on deep sclerectomy, and another 17 in the trabeculectomy paper. Patients were aged 56.9 and 48.1 years, respectively. In the deep sclerectomy group, the procedure was successful in 88% of the patients, compared to 65% in the trabeculectomy group with a follow-up of 42 and 52 months [18, 20].
As reported by others, aphakic children had a higher failure rate after glaucoma surgery. In a group of patients with different entities of juvenile glaucoma, the failure rate after trabeculectomy with MMC in aphakic children was as high as 60% [21]. Another study analysed the outcome after goniotomies in chronic childhood uveitis. The success rate was only 36% (4 eyes) in aphakic patients, while 86% (25 eyes) of the procedures were successful in the phakic group [11]. In our retrospective setting, there has been a bias towards a certain surgical technique, which also results in uneven balance of our two groups. The aphakic children in this series were treated more frequently with the modified deep sclerectomy in order to avoid vitreous in the trephination. We are aware that comparison of the two reported groups is therefore limited in some aspects but still allows comparison of different aspects of these two surgical procedures.
One prospective randomized trial exists that compares these two different surgical techniques in adults with primary open angle or pseudoexfoliation glaucoma [22]. In this study, both techniques provided sufficient IOP reduction without any statistical significant difference. In another study by the same authors, trabeculectomy appeared to be more suitable for higher IOP levels and longer life expectancies than deep sclerectomies [23]. Our results regarding IOP reduction in juvenile uveitic glaucoma are in accordance with these previous observations, as trabeculectomy with MMC was more effective than modified deep sclerectomy with MMC.
The overall rate of additional surgery including limited postoperative interventions and additional glaucoma surgery was similar in both groups. In the mdS group, four additional glaucoma procedures had to be performed compared to one in the TE group, while in TE group, eight patients needed limited surgical interventions to adjust IOP to levels between 10 and 20, including downsizing of the filtering bleb, suture lysis and needling procedure. Postoperative choroidal detachment and a shallow anterior chamber occurred in as many as 50% of the patients with TE. This hypotony is probably due to reduced aqueous production and increased uveoscleral drainage because of postoperative inflammation. It can be speculated that these complications might have accelerated the development of posterior synechies, cataract formation and macular abnormalities. Cataract surgery was necessary in 5 of 11 phakic children in the TE group during follow-up (mdS group: 0 of 1 phakic patient required this treatment), but it is impossible to say whether the cataract formation was induced through surgery or inflammation. The vitreous prolapse that occurred during the immediate postoperative period in two aphakic eyes was unfortunate. However, after surgical removal of the vitreous strands, the trephination hole was overt.
In summary, glaucoma surgery in children with secondary uveitic glaucoma still remains difficult. Our data suggest that IOP can be sufficiently reduced using standard trabeculectomy with MMC and deep sclerectomy with MMC, but TE with MMC may be more effective. However, additional surgeries to adjust IOP finally were common for both groups. For aphakic children, the modified sclerectomy described earlier appears to be a good technique for avoiding vitreous prolapse.