Aims: The aim of the ab externo trabeculectomy (AET) is to remove the external portion of the trabecular meshwork (ETM) responsible for the main aqueous outflow resistance in glaucoma patients, with no opening of the anterior chamber. ETM characteristics were evaluated with a confocal microscope. Methods: A prospective comparative observational case series was performed in 60 consecutive medically treated patients with primary open angle glaucoma and eight postmortem normal donors' eyes that underwent AET. Once deroofing the Schlemm' s canal (SC), a deeper dissection led to removal of a coherent membrane (ETM) which allowed satisfactory aqueous egress through the remaining intact internal trabecular meshwork (TM) layers. After fixation with acetone and immunostaining with anti-vimentin antibody, ETM were analysed with a confocal microscope. Results: Glaucomatous ETM (mean thickness: 29.5 (7.6) µm) were characterised by a severe paucicellularity compared with the controls (respectively 37.3 (9.7) cells/area and 167.5 (24.9) cells/area, p<10 −4 ). ETM analysis showed involvement of both cribriform and corneoscleral layers. ETM cell density was significantly decreased in case of preoperative fluorometholone instillation. Conclusion: Paucicellularity of glaucomatous TM is confirmed by this original technique. Structural characteristics of the ETM, whose removal allows satisfactory aqueous egress, suggest that aqueous outflow resistance not only involves inner wall of SC and juxtacanalicular meshwork but also corneoscleral trabecular layers.

Corneal wound healing often leads to the development of scar tissue with loss of transparency. Reconstitution of transparent corneal stroma depends on the regulation of the biosynthetic activities of postlesional keratocytes and also to a large extent on the limitation of matrix degradation, attributed essentially to the upregulation of matrix metalloproteases and especially MMP-9. Using a standardized method for the production of reproducible corneal lesions by burning with iodine vapors, we could show that the local application of 0.5 mg/ml L-fucose reduced significantly MMP-9 upregulation and accelerated the recovery of the epithelial layer of the cornea. The iodine vapor used in the experiments produces a rapid loss of epithelium with no or slight effect below the basement membrane. A relatively rapid regrowth of epithelium was observed. The speed of this reepithelialization was stimulated by the local application of fucose. At 48 h after burn, there was a difference between fucose-treated and control corneas (epithelial thickness was about 50 µm for fucose-treated corneas and 37 µm for control corneas). Culture media of in vivo fucose-treated corneas showed an important decrease of MMP-9 activity (–51%, n = 6, p < 0.01). It appears that the in vivo fucose treatment reduced the MMP-9 activity released in the media. This effect is significant 24 h after iodine vapor burn. In order to study the effect of fucose on normal corneas, it was added to rabbit as well as human cornea explant cultures, and the production and release of MMP-9 was determined by zymography. Fucose at a concentration of 0.5 mg/ml produced a 70% decrease of MMP-9 activity released in the medium by corneal explant cultures. Other mono- and oligosaccharides were also tested. Besides lactose, fucose-rich oligosaccharides also produced significant inhibition. Galactose, melibiose, mannose and glucose were inactive. These results justify the use of fucose for the local treatment of corneal wounds.