11
treated topically 3 months postoperatively. One eye healed completely, while the other eye
developed endophthalmitis. We have insufficient data regarding the third patient as she
presented and was treated in another country, 4 months postoperatively. On return to our
center she had no evidence of prior endophthalmitis. Given her excellent outcome with no
sequela, we cannot rule out the possibility that the diagnosis of endophthalmitis in this patient
may have been incorrect. The affected eye in the other two patients progressed to phthisis.
Tube retraction was one of the late complications, encountered in only three patients,
after an average follow up period of 15.5 ± 4.8 months. This resulted in tube blockage that
required tube reinsertion (without handling the plate) in 2 cases and a new Ahmed
implantation in one case. One of the cases occurred early in our series when absorbable
suture had been tried to fixate the plate. At surgery, the suture was found to be dissolved and
the plate had migrated posteriorly.
Seven eyes lost more than 2 lines of Snellen visual acuity. Two had phthisis due to
endophthalmitis and in another two visual loss was related to hypotony. Three eyes, all
suffering from uveitis, lost vision for reasons not related to the implant (posterior sub-
capsular cataract, inflammatory optic neuritis, and intractable uveitis caused by an intraocular
lens placed at another center). These 3 eyes were not considered failures in our statistical
analysis.
