Statistics from Altmetric.com
Editor,—The use of postoperative subconjunctival 5-fluorouracil (5-FU) injection to inhibit scarring after glaucoma filtering surgery is now commonplace. Uncomplicated subconjunctival 5-FU injection is unlikely to produce endothelial toxicity, as post-injection anterior chamber drug levels remain low.1However, inadvertent intraocular exposure can be toxic to the cornea.2 We report a case of postoperative subconjunctival 5-FU injection with accidental passage of the drug directly into the anterior chamber.
A left trabeculectomy was performed on a 79 year old white woman using peroperative 5-FU (25 mg/ml). Two days later the patient rubbed her eye and the anterior chamber became transiently flat. Over the next 4 weeks the drainage bleb encysted, and the patient received 5 mg (0.2 ml) 5-FU into the superior bulbar conjunctiva. Immediately after injection the intraocular pressure (IOP) dropped from 23 to 10 mm Hg, and air bubbles from the syringe were seen in the anterior chamber. It was clear that 5-FU had gained direct intraocular access. The anterior chamber was washed out with balanced salt solution within 30 minutes of exposure. The next day the IOP measured 15 mm Hg and the cornea was clear. Two uneventful subconjunctival injections of 5-FU (5 mg) were subsequently administered. Eleven months after her glaucoma surgery the patient underwent endocapsular cataract extraction and lens implant with one further peroperative subconjunctival 5-FU (5 mg) injection.
One month after the initial glaucoma surgery corneal specular microscopy was normal, although patient photophobia prevented detailed assessment. A second specular micrograph performed before cataract surgery revealed a central cell density of 2390 cells/mm2, an average cell size of 430 μm2, and a cell variation of 40%. Two years after the glaucoma drainage operation a third specular micrograph revealed a left eye central cell density of 1887 cells/mm2 (right eye 2238 cells/mm2), an average cell size of 530 μm2 (right eye 447 μm2) and cell variation of 44% (right eye 41%). Endothelial polymegathism was noted (Fig 1B). The corneal thickness measured 0.46 μm in the left eye and 0.42 μm in the right. The corrected visual acuity was 6/9 and the IOP 18 mm Hg on no therapy.
There are two possible mechanisms by which 5-FU may compromise cellular function. The first is inhibition of cell replication, and the second a toxic effect, either by direct drug action or an alkali effect (pH 8.9). Corneal endothelial cell toxicity has been reported in in vitro animal studies.3-5 However, the in vitro model does not allow for aqueous dilution, aqueous turnover, or aqueous pH buffer effects. In addition, corneal endothelial cells replicate in animal models, unlike their human counterparts. Direct and/or alkali toxicity is the more probable mechanism in the human cornea.
Two other cases of 5-FU gaining direct access to the anterior chamber have been reported in humans.2 The authors described the use of subconjunctival 5-FU (50 mg/ml) after a bleb needling procedure. Anterior chamber washout was not performed and severe corneal oedema developed at day 1 which resolved completely after 6 months. Increased endothelial cell pleomorphism was noted.
In our case report, the significant fall in intraocular pressure after injection suggests that the needle entered the bleb cyst, increased intracyst pressure, and enabled 5-FU to enter the anterior chamber.
The first and second specular micrograph findings indicate that the endothelium was relatively undamaged by 5-FU exposure, and the drop in cell count noted in the third micrograph is compatible with published reports of cell loss encountered after endocapsular cataract surgery. In conclusion, we did not experience clinically significant endothelial cell toxicity, presumably due to the dilution and pH buffering effects of aqueous, followed by prompt anterior chamber washout. As 5-FU is potentially toxic in the eye ,2 our experience suggests that care should be taken when injecting around an encysted bleb.