Statistics from Altmetric.com
Editor,—It is well established that early intravitreal antibiotics are the treatment of choice for bacterial endophthalmitis1 but the choice of antibiotics remains controversial.2-5 Most centres rely on combination therapy and aminoglycosides are often used to treat Gram negative organisms. Aminoglycosides have a narrow therapeutic index and therefore a relatively small increase in concentration can cause significant toxicity. Following concerns about the risks of retinal toxicity caused by intravitreal gentamicin many ophthalmologists now use 0.4 mg intravitreal amikacin even though its use has also been associated with retinal toxicity.2 3We report a case of retinal toxicity in a patient treated with intravitreal amikacin for postoperative bacterial endophthalmitis. This case serves to remind ophthalmologists of the risk of amikacin therapy and presents an argument for an alternative strategy in the treatment of bacterial endophthalmitis.
A 60 year old highly myopic woman presented with an acute history of increasing ocular pain and redness 5 days after routine phacoemulsification cataract extraction and intraocular lens implant. She had no other ocular or systemic history.
Initial examination revealed a visual acuity of 6/18, significant anterior chamber inflammation with a fibrin deposit over the pupil. A diagnosis of bacterial endophthalmitis was made and the patient was commenced on intravenous ciprofloxacin and hourly gentamicin, vancomycin, and dexamethasone eye drops. A vitreous biopsy was performed and 2 mg of intravitreal vancomycin (0.1 ml of 20 mg/ml) and 0.4 mg amikacin (0.1 ml of 4 mg/ml) were administered. The intravitreal antibiotics were prepared by a senior pharmacist using a typewritten protocol. The following day vision had deteriorated to no perception of light and the patient had developed a relative afferent pupillary defect. Microscopy and culture failed to demonstrate any causative organism. The patient was started on 40 mg oral prednisolone but failed to manifest any improvement in acuity. Subsequently she was transferred to a specialist vitreoretinal unit where a diagnosis of retinal amikacin toxicity was made. Fluorescein angiogram taken 16 days after initial presentation confirmed the presence of retinal haemorrhage and a discrete area of capillary non-perfusion with significant late fluorescein leakage (Figs 1 and2).
Intravitreal aminoglycosides are still recommended for use in the treatment of bacterial endophthalmitis6 7 despite the risk of retinal toxicity from intravitreal gentamicin and reports that amikacin in doses of 0.2–0.4 mg can also cause toxicity.2 3 The low incidence of bacterial endophthalmitis means that physicians and pharmacists do not routinely prepare intravitreal antibiotics. Because of the low therapeutic index of aminoglycosides even relatively small dilution errors can cause retinal toxicity. For these reasons it is our practice to use a third generation cephalosporin (0.1 ml of ceftazidime 20 mg/ml) instead of an aminoglycoside. Primate studies have shown that intravitreal ceftazidime is non-toxic in doses of 2.25 mg8 and ceftazidime has also been found to be effective against the Gram negative organisms responsible for postoperative endophthalmitis.9 Over the past 18 months we have used intravitreal ceftazidime (2 mg) and vancomycin (2 mg) in 21 patients with good therapeutic effect and with no adverse reactions. Owing to the risk of severe retinal toxicity with intra- vitreal aminoglycosides we recommend the use of ceftazidime instead of amikacin or gentamicin.