Main

Sir,

Endophthalmitis is a well-recognised late complication of glaucoma filtration surgery. The causative organisms differ from those of acute postoperative endophthalmitis. We present the first case description of late bleb-related endophthalmitis caused by Streptococcus agalactiae or group B Streptococcus (GBS).

Case report

A nondiabetic 86-year-old woman presented to the emergency eye service in 2001 with a 1-day history of painful, inflamed left eye associated with marked reduction in visual acuity. She was known to have bilateral advanced primary open-angle glaucoma, which had been treated with bilateral trabeculectomies without antifibrotic agents (right 1998, left 1999), and she had been bilaterally pseudophakic for 12 months. Recently, she had undergone treatment for right cystoid macular oedema. At 2 weeks before presentation, her visual acuities were 6/36 (right) and 6/24 (left), and a slow leak was detected in her left conjunctival surgical bleb. The anterior chambers were deep.

At presentation, her visual acuities were counting fingers (right) and perception of light (left). The left conjunctiva was severely injected. The bleb was opaque with marked corneal oedema and intense fibrino-cellular reaction in the anterior chamber. The intraocular pressures were 12 mmHg (right) and 43 mmHg (left). A diagnosis of left bleb-related endophthalmitis was made. Vitreous humour was sampled from this eye and ceftazidime 2 mg and vancomycin 1 mg were injected intravitreally. Topical postoperative treatment consisted of cefuroxime 5% hourly, gentamicin 1.5% hourly, atropine 1% tds, timolol 0.5% b.i.d., and prednisolone acetate 1% 2-hourly. Oral ciprofloxacin 750 mg b.i.d. and acetazolamide SR 250 mg qd were also prescribed. Vitrectomy was not performed because of its uncertain benefit in bleb-related endophthalmitis.

After 24 h, GBS was isolated from the vitreous fluid. Topical gentamicin and oral ciprofloxacin were replaced by topical penicillin hourly and intravenous imipenem 1 g t.i.d. However, after 2 days there was no improvement and the vitreous tap was repeated. Vancomycin 1 mg, ceftazidime 2 mg, and dexamethasone 0.4 mg were injected intravitreally. This second sampling yielded GBS on culture again.

At 4 days after presentation, the left eye had no perception of light. The conjunctival bleb had perforated and the cornea was opaque. After 1 month, the eye was phthisical.

Comment

GBS is a well-documented neonatal pathogen, although invasive adult GBS disease is also increasing in frequency. However, GBS remains an uncommon cause of intraocular infections. Endogenous GBS endophthalmitis in adults is usually secondary to meningitis or endocarditis,1 while endophthalmitis in neonates results from septicaemia.2 Exogenous intraocular infection with GBS is uncommon and occurs mostly in patients with severely damaged corneal surfaces.3 A few cases of early postoperative endophthalmitis (postpenetrating keratoplasty or cataract surgery) because of GBS have also been reported.3,4 Delayed onset endophthalmitis is a well-recognised complication of conjunctival filtration blebs. Although most series report a preponderance of streptococcal species as the causative organisms, these are usually viridans streptococci or Streptococcus pneumoniae.5 To our knowledge, this is the first description of late onset GBS infection associated with a conjunctival filtration bleb. In this case, there was no other focus of infection with GBS and blood cultures were negative.

The optimal therapy for ocular infections with GBS remains uncertain. Intravitreal administration of ceftazidime and vancomycin achieves antibiotic levels that exceeds the minimum inhibitory concentrations for the organism. The Endophthalmitis Vitrectomy Study concluded that treatment with systemic antibiotics conferred no additional benefit in the management of postcataract endophthalmitis,6 although there was a role for vitrectomy. However, the extrapolation of these results to other categories of endophthalmitis remains a topic of debate.7 Intravitreal antibiotics may only maintain effective local concentrations for periods of 36–48 h8 and leakage of aqueous humour through the perforated conjunctival bleb in this patient may have contributed to persistence of the organism in samples taken 48 h following initial intravitreal therapy. GBS remains universally susceptible to penicillin.9 However, most penicillins penetrate poorly into the vitreous humour. Similarly, intraocular levels of systemic vancomycin10 and cephalosporins are variable. Although oral ciprofloxacin achieves adequate levels in vitreous humour for the treatment of Gram-negative infections, the concentrations achieved may be subinhibitory for most Gram-positive organisms.11 Systemic imipenem therapy achieves effective penetration into ocular fluids.12 Hence it was added to this patient's therapy.

The outcome of GBS endophthalmitis appears to be poor. Late infection in this patient resulted in rapid and complete loss of vision. In earlier reported case studies,3 three out of the four cases of early postoperative GBS endophthalmitis resulted in a blind eye. Similarly, a case series of endogenous GBS endophthalmitis also reported poor outcomes.1 Perhaps, bleb revision immediately upon detection of a leak would prevent endophthalmitis in some exogenous cases.