A 53-year-old white male who worked in a sawmill was referred to the Spokane Eye Clinic in Washington with a 1-week history of hyperemia and complaints of a foreign body sensation in his right eye. Although the patient was using moxifloxacin q.i.d. as prescribed by his local optometrist, he reported increasing pain over time.
The patient had a complex ocular history that was significant for bilateral trabeculectomies in 1996. He subsequently developed endophthalmitis in his left eye that was attributed to Streptococcus bacteria. Despite aggressive treatment with a vitrectomy and intraocular antibiotics, the patient developed a retinal detachment and completely lost the vision in his left eye.
Upon initial examination at the Spokane Eye Clinic, the patient's vision was 20/200 OD, and he had no light perception in his left eye. The pale, ulcerated bleb in his right eye was leaking, and his IOP measured 5 mm Hg. The conjunctiva surrounding the bleb was injected. A slit-lamp examination showed 2+ cells in the anterior chamber, but no vitreous cells were noted through the moderate nuclear sclerotic cataract in the left eye.
The patient was diagnosed with trabeculectomy-related endophthalmitis. My colleagues and I immediately performed a vitrectomy and administered intravitreal vancomycin, ceftazidime, and dexamethasone. In addition to following the patient closely, we treated him with topical moxifloxacin q2h, prednisolone acetate q.i.d., atropine q.d., and 500 mg of oral levofloxacin q.d. Aqueous and vitreous cultures obtained during the vitrectomy were negative for bacteria and fungus.
Over the next several days, the inflammation and pain in the patient's right eye decreased. Because the large bleb was still leaking, however, my partner excised the necrotic tissue and sutured a free conjunctival graft harvested from the patient's fellow eye onto the healthy, vascularized conjunctiva of his right eye. Although the infection improved and his vision increased to 20/80 OD postoperatively, his IOP began to rise as scar tissue developed around the bleb. We discontinued the antibiotics and treatment with topical hypotensive medications. Because the patient appeared to be responding to treatment, we did not perform additional tests to identify the organism responsible for his infection.
Two weeks after the revision of his bleb, the patient returned with a complaint of renewed eye pain. An examination of his right eye showed hyperemia and a white infiltrate in the new bleb. Surface cultures have been shown to have limited efficacy in identifying organisms in blebitis. For that reason, we did not obtain conjunctival cultures and instead placed the patient on empiric therapy with fourth-generation fluoroquinolones. Despite intensive treatment with topical moxifloxacin and gatifloxacin, the bleb developed a small leak. Within a few days, the leak progressed to a full-thickness melt of the conjunctiva and sclera that revealed the underlying uveal tissue.
I immediately took the patient to the OR, excised the necrotic conjunctival and scleral tissue, and fashioned a new trabeculectomy flap by suturing a pericardial graft to the remaining viable sclera. I replaced the tissue of the bleb with a sliding conjunctival graft. The patient's symptoms initially improved with the administration of topical fourth-generation fluoroquinolones that were gradually tapered over the next several weeks.
Two weeks after the placement of the pericardial graft, the patient's pain and hyperemia returned. My colleagues and I noted cells in the anterior chamber at this time. The patient's symptoms improved on a regimen of oral/topical fluoroquinolones and topical fortified tobramycin. After his medications were tapered, however, the symptoms of infection recurred.
HOW WOULD YOU PROCEED?
- Would you recommend another vitrectomy and treatment with intraocular antibiotics?
- Continue to treat for all possible infectious organisms?
- Treat the patient's inflammation with prednisone?
CLINICAL AND SURGICAL COURSE
Due to the recurrence of the patient's symptoms, we suspected that previous antibiotic therapy had not fully eradicated the underlying, unidentified infectious organism. The patient underwent another vitrectomy to his right eye and received intravitreal injections of vancomycin, amikacin, and dexamethasone. The cultures collected during the vitrectomy were again negative for bacteria and fungus.
Three days postoperatively, we observed a whitish, flocculent material in the superior chamber near the sclerotomy. Under a gonioscopic view in the OR, we aspirated some of the material from the anterior chamber and injected more antibiotics into both chambers of the eye.
Cultures of the material aspirated from the anterior chamber of the patient's right eye grew Aspergillus fumigatus. Although my colleagues and I were concerned that the sample could be contaminated by an outside source, we treated the patient with intravenous and intravitreal antibiotics (including voriconazole) as well as subconjunctival amphotericin B.
At this time, the patient's vision measured a bare 20/200, and his IOP was elevated to 35 mm Hg. A slit-lamp examination showed an irregular corneal epithelium, stromal edema, and the presence of flocculent material in the superior part of the anterior chamber near the trabeculectomy site. The etiology of the recurrent inflammation remained uncertain, and we were concerned that it could represent a reaction to the pericardial graft or an autoimmune condition such as scleritis. Further examination revealed an incipient scleral staphyloma superiorly.
Consultants from another institution reviewed the patient's case and suggested that his inflammatory symptoms could be partially due to medicamentosa. The patients' symptoms gradually improved after his medications were tapered, but they recurred a few weeks later and did not resolve after additional treatment with topical moxifloxacin and prednisolone acetate. We prescribed 60 mg of oral prednisone and monitored him closely due to our concerns that the inflammation in his right eye was caused by autoimmune uveitis/scleritis. The patient's pain improved on this regimen, but an examination showed an increase in the amount of purulent debris and 4+ cells in the anterior chamber as well as a hypopyon. The patient was once again treated with a vitreous tap and intravitreal antibiotics. The cultures were again negative for bacteria or fungi.
We referred the patient to another institution for further evaluation. After examining the patient, the consulting surgeons decided to excise scleral tissue superior to and underlying the existing bleb. They also removed corneal tissue adjacent to the bleb during the same procedure and covered the area with donated scleral and conjunctival grafts. An analysis of the excised scleral tissue confirmed the diagnosis of Aspergillus infection.
Postoperatively, the patient was placed on amphotericin B (topical and subconjunctival injection [0.5 mg/injection up to two times a week as tolerated]) and oral voriconazole 200 mg b.i.d. When routine blood tests showed that the patient's liver enzymes were elevated due to the systemic antifungal treatment, the voriconazole was discontinued. For the next several months, he was treated with intravenous caspofungin (loading dose, 75 mg; maintenance dose, 50 mg q.d.) and topical amphotericin B.
OUTCOME
After 4 months, the patient's bleb-related endophthalmitis was successfully eradicated by the the antifungal regimen described herein (Figure 1). Nine months after his initial presentation, he underwent cataract extraction and the placement of an Ahmed Glaucoma Valve (New World Medical, Inc., Rancho Cucamonga, CA). His postoperative BCVA was 20/200, and his IOP remained in the upper teens with Cosopt (Merck & Co., Inc., West Point, PA), Travatan (Alcon Laboratories, Inc., Fort Worth, TX), and Alphagan (Allergan, Inc., Irvine, CA).
DISCUSSION
Endophthalmitis after trabeculectomy is a well-known and much-dreaded complication of filtering surgery. Because bacterial endophthalmitis that occurs after penetrating glaucoma surgery tends to have worse outcomes and is caused by a different spectrum of pathogens than similar infections associated with cataract extraction,1,2 the recommendations from the Endophthalmitis Vitrectomy Study do not apply to the late-onset case described in this article.2 The standard treatment for bleb-related endophthalmitis is therefore vitrectomy-and-inject rather than tap-and-inject, regardless of the patient's visual acuity at presentation.
The patient described herein was treated with vitrectomy-and-inject after his first evaluation due to the diagnosis of trabeculectomy-related endophthalmitis. Regrettably, the results of his initial culture were negative and therefore provided no basis for altering his therapy. Negative vitreous cultures are not uncommon in cases of bleb-related endophthalmitis.3 Compared with patients who have positive cultures, those with negative cultures have a better prognosis. Busbee et al suggest that a negative culture may reflect a lower bacterial load or infection with less aggressive organisms.2 Previous therapeutic regimens could also influence the results of cultures and cause false negative readings.
In this case, the patient appeared to respond well to therapy but relapsed as the antibiotic regimen was tapered. This pattern repeated itself several times. My colleagues and I did not identify A. fumigatus as the source of infection until we obtained a third set of cultures. Even then, the positive result was treated with skepticism, because A. fumigatus is frequently found in contaminated laboratory cultures. Furthermore, the literature reports that infection with Aspergillus species is difficult to diagnose in cases of scleritis4 and postcataract endophthalmitis.5
Our patient's diagnosis was partially hampered by his positive response to antibiotic therapy, which included several courses of topical fourth-generation fluoroquinolones. Studies have shown that commercial topical preparations of moxifloxacin and gatifloxacin have in vitro antifungal activity against ocular Candida species6 as well as clinical efficacy against fungal keratitis.7 The recurrence of the patient's infection after treatment with topical fluoroquinolones suggests that these agents partially suppressed the underlying fungal pathogen. Because the fungus had invaded the sclera, however, as evidenced by the recurrent scleral melting, the topical agents were not sufficiently powerful to eradicate the organism. Overall, the patient's apparent partial response to antibiotic therapy delayed his eventual diagnosis and the initiation of definitive treatment.
Aspergillus molds are ubiquitous and are particularly associated with decaying vegetation. Although these molds are not generally pathogenic (except in immunosuppressed patients), their introduction into the eye by trauma or surgery can cause endophthalmitis. The poorly vascularized tissue of filtering blebs and the underlying sclera may be particularly susceptible to invasion, especially if the eye has previously been exposed to antifibrotic agents. Fortunately, Aspergillus is sensitive to amphotericin B, voriconazole, and caspofungin, all of which were used to treat this patient. Individuals treated with these agents must be monitored for toxicity, however, as evidenced by the elevated liver enzymes observed in the patient described herein.
Barbara Smit, MD, PhD, is a glaucoma consultant at the Spokane Eye Clinic and a clinical instructor for the University of Washington School of Medicine in Washington. She acknowledged no financial interest in the products or companies mentioned herein. Dr. Smit may be reached at (509) 456-0107; bsmit@spokaneeye.com.
