Peribulbar fungal abscess and endophthalmitis following posterior subtenon injection of triamcinolone acetonide

Abstract

A58-year-old man was referred for treatment of diabetic retinopathy. His best corrected visual acuity (VA) was 0.07 OD and 0.02 OS. He agreed to our recommendation that the macular oedema in the left eye be treated by a subtenon triamcinolone acetonide (TA) injection. The eye was anaesthetized with 4% topical lidocaine, and disinfected along with the lid margins and skin with 5% povidone iodine. A lid speculum was used, and the superonasal conjunctiva was incised to inject 20 mg of TA into the posterior subtenon space using a 27-gauge cannula with a dull tip. Topical ofloxacin ointment was instilled and levofloxacin q.i.d. was prescribed. Three months later, the patient reported diplopia and deep ocular pain OS. He had ptosis and restricted motility, and VA had dropped to 0.5 OS. The anterior chamber was quiet and the conjunctiva was not injected. Magnetic resonance imaging (MRI) showed a high-intensity lesion in the superonasal peribulbar area (Fig. 1A). Two months later, the diplopia worsened, and an epibulbar abscess was found superonasally (Fig. 1B). A repeat MRI revealed exophthalmos with two dome-like periocular cysts OS (Fig. 1C). Photograph of left conjunctiva, fundus photograph and magnetic resonance images (MRI) of the brain and orbit. (A) An MRI T1 weighting image (T1WI) performed 12 weeks after a posterior subtenon injection of triamcinolone (TA) showing a high-intensity lesion in the superonasal peribulbar area. (B) Two months later, bulbar conjunctiva is oedematous, and an abscess can be seen at the superonasal region that corresponds to the injection site. (C) T2WI performed at the same time shows left exophthalmia and two, dome-like retrobulbar cysts showing low intensity in T1WIs and high intensity in T2WIs. (D) After surgical debridement of the abscess, the conjunctival oedema has disappeared and no abscess is seen. (E, F) Repeat MRI performed 4 weeks afterwards show a localized, high-intensity retrobulbar lesion in the T1WI in the superonasal region. (G) Fundus photographs taken at the same time show diabetic retinopathy with panfundus photocoagulation scars, vitreous opacities, periphlebitis at the superior arcade vessels, pale optic disc, and macular retinal detachment. * indicates a more peripheral site. (H) Localized retinal haemorrhages and a whitish subretinal mass of 2 × 3 disc diameters in size can be seen in the superior peripheral site indicated by * in (G), which was probably associated with the retinal detachment and periphlebitis. Note that the picture is inverted because it was taken using a slit-lamp biomicroscope and a 90-D lens. (I) Optical coherence tomogram showing macular oedema and retinal detachment. The abscess was incised to drain a considerable amount of yellowish-white purulent fluid. Gatifloxacin, pimaricin, levofloxacin (400 mg/day) and itraconazole (100 mg/day) were given topically and fluconazole (100 mg/day) intravenously. Cultures yielded Scedosporium apiospermum, which was confirmed by DNA analysis. Therefore, itraconazole was replaced with voriconazole (300 mg/day), to which the fungus was sensitive. The symptoms gradually decreased (Fig. 1D), but 4 weeks later the patient developed a central scotoma and a concentric visual field defect OS. His VA had decreased to 0.02. Repeat MRI showed a retrobulbar lesion in the superonasal region (Fig. 1E, F). Extensive vitritis with opacities, pale optic disc, periphlebitis, serous detachment of the macula, focal retinal haemorrhages and a whitish subretinal peripheral mass were noted (Fig. 1G–I). The former abscess, now replaced by fibrosis, was irrigated with balanced salt solution containing voriconazole, amphotericin B and ceftazidime. Six days later, pars plana vitrectomy and cataract extraction were performed. The mass gradually decreased, the retina was reattached, and VA improved to 0.5. No recurrence was observed in the following 6 months. Triamcinolone acetonide is a long-acting steroid, and intravitreal or subtenon injection of TA has been used to treat macular diseases (Martidis et al. 2002). However, an intravitreal injection can cause a retinal detachment and endophthalmitis, and a subtenon TA injection can cause secondary ptosis and ocular infections (Engelman et al. 2004; Erol & Topbas 2006;Kusaka et al. 2007; Oh et al. 2007). A Medline search identified three patients with an orbital abscess or endophthalmitis following a subtenon TA injection (Engelman et al. 2004; Kusaka et al. 2007; Oh et al. 2007). One of these was caused by Pseudallescheria boydii and did not respond to intensive itraconazole therapy, leading to phthisis (Oh et al. 2007). The anamorph, S. apiospermum, is a filamentous fungus found in soil and polluted water. Its teleomorph is P. boydii, an emerging opportunistic fungus that usually affects immunocompromised hosts. Although the possibility of endogenous origin in the diabetes patient cannot be fully ruled out, the source of P. boydii in our case may have been direct contact with polluted water or rubbing the eye with contaminated fingers. which allowed the organism to gain access to the subtenon space via the conjunctival incision. The organisms may have passed through the eye wall to cause the endophthalmitis. Endophthalmitis caused by P. boydii has often required enucleation or evisceration, and can even be fatal (McGuire et al. 1991). In our patient, the fungus was sensitive to voriconazole and intraocular transmission of the drug was increased after vitrectomy, which contributed to a favourable outcome. Subtenon TA injection is often performed on an immunocompromised patient. Ophthalmologists should recognize that a subtenon TA injection can cause orbital infections and endophthalmitis, especially fungal infections in compromised patients. In diabetes, both immunosuppression and high glucose levels promote fungal infection. A comprehensive search for a fungus should be made without delay when symptoms develop, because swift antifungal therapy combined with vitrectomy is most likely to be effective. Increasing frequencies of subtenon TA injection would increase the incidence of severe ocular infection, and a systematic survey, such as that carried out for cataract surgery (Oshika et al. 2007), would help to establish the proper prophylactic and therapeutic paradigm. This study was supported by funds from Researches on Sensory and Communicative Disorders. Ministry of Health, Labour and Welfare, Japan. None of the authors has any proprietary interest in any material or method mentioned. The authors would like to thank Tadao Nakano for his extremely important contributions concerning microbiological examinations.