Purpose
To report the clinical settings, microbiological isolates, and best-corrected visual acuities (BCVA) of patients with persistently culture-positive exogenous fungal endophthlamitis.
Design
Retrospective consecutive case series.
Methods
Setting : Tertiary referral center. Patient Population : Sixteen eyes of 16 patients with at least 2 consecutive positive vitreous cultures between 1981 and 2015. Interventions : Intravitreal antifungal injection, pars plana vitrectomy (PPV). Main Outcome Measure : Clinical settings, microbiologic isolates, BCVA.
Results
The most common clinical settings were after cataract surgery (9/16, 56%), glaucoma surgery (4/16, 25%), and trauma (2/16, 13%). The most common single fungal isolate was Candida (4/16, 25%), but 75% of all isolates were molds. Treatment for presumed bacterial endophthalmitis was given initially in 14 patients (88%). All patients underwent a vitrectomy during the course of their treatment, and all received intravitreal or systemic antifungal therapy. The mean initial BCVA was 1.76 ± 0.9 logMAR (Snellen equivalent ≈20/1200), and the mean final BCVA was 1.84 ± 1.2 logMAR (≈20/1400, P = .83). The 9 patients (56%) who had intraocular lens (IOL) and capsular bag removals had better final BCVAs than those who did not ( P = .011). The BCVAs were similar in eyes with yeast and mold ( P = .37). The visual acuity at the last follow-up was ≥20/40 in 13% (2/16), ≥20/400 in 50% (8/16), and no light perception in 25% (4/16).
Conclusions
Candida was the single most common isolate, but the majority of isolates were molds. Eyes managed with PPV and removal of the IOL and capsular bag had better visual outcomes. Persistently culture-positive fungal endophthalmitis was associated with poor final visual acuities.
Exogenous fungal endophthalmitis can occur from direct inoculation after surgery or trauma or extension from infectious keratitis. The incidence of fungal endophthalmitis is approximately 1%–4% after trauma, 0.002%–0.005% after cataract surgery, and less than 0.5% after infectious keratitis. After trabeculectomies, the 5-year cumulative incidence of culture-proven endophthalmitis is approximately 0.5%; however, only 1% is fungal. Fungal endophthalmitis after intravitreal injections is rare and was primarily associated with contaminated triamcinolone. The symptoms of fungal endophthalmitis can occur acutely or insidiously, with chronic inflammation and infection leading to significant damage to intraocular structures.
The purpose of the study is to identify the clinical settings, fungal isolates, and best-corrected visual acuities (BCVA) of patients with persistently vitreous culture–positive exogenous fungal endophthalmitis.
Methods
The retrospective consecutive case series was approved by the Institutional Review Board of the University of Miami Miller School of Medicine and was compliant with the Health Insurance Portability and Accountability Act of 1996. The research adhered to the tenets of the Declaration of Helsinki. The inclusion criterias were patients with the same exogenous fungal organism identified on at least 2 consecutive vitreous cultures obtained on separate days at the Bascom Palmer Eye Institute from January 1, 1981 to December 31, 2015. Patients with polymicrobial cultures, endogenous fungal endophthalmitis, and incomplete medical records were excluded.
The vitreous samples were plated on Sabouraud agar and incubated at 35°C for 72 hours, then examined daily for the growth of fungal colonies for 2 weeks. Positive samples were stained with Giemsa or calcofluor white. Fungal identification and antifungal susceptibilities were determined by sending the samples to the Fungus Testing Laboratory in San Antonio, Texas. The fungal sensitivities were not repeated for the second set of positive vitreous cultures.
The treatment regimen was determined by the physician based on the individual patient’s clinical course. There were no predefined protocols for the timing and types of treatments. The indications for re-treatment were clinically persistent or worsening endophthalmitis, as evidenced by increasing fibrin or hypopyon, persistent fungal infiltrates, worsening visual acuity, and/or persistent intraocular inflammation.
Statistical calculations were performed using the Statistical Package for the Social Sciences software (SPSS Inc, Chicago, Illinois, USA), with a P value less than .05 being considered statistically significant. Snellen visual acuity was converted to its logarithm of minimal angle of resolution (logMAR) equivalent, with counting fingers being assigned a value of 1.9, hand motion 2.3, light perception 2.7, and no light perception 3.0. BCVA is presented as the mean logMAR ± standard deviation, followed by the approximate Snellen chart equivalent. The visual acuities were analyzed using Student t -test and one-way analysis of variance with Tukey post hoc analyses.
Results
Of the 165 patients with positive fungal vitreous cultures over the 35-year study period, there were 18 patients (11%) with 2 positive fungal cultures of the same organism. Two cases from the 1980s were unavailable for review; therefore, 16 eyes of 16 patients were included in the study. The mean age was 68.6 ± 18 years; 10 patients were male (63%), and half of the eyes were right(8/16). The most common past medical histories were hypertension (7/16, 44%) and diabetes mellitus (4/16, 25%). Excluding the 4 diabetics, no one was systemically immunocompromised. Four patients (25%) were on topical steroids on presentation.
The most common past ocular histories were glaucoma (5/16, 31%), corneal transplants (2/16, 13%), and retinal detachment (1/16, 6%). The mean follow-up period was 42 months (range: 7–288 months). Table 1 summarizes the patient demographics.
| Result | |
|---|---|
| Mean age | 68.8 ± 18 years |
| Mean follow-up | 42.0 ± 72 months |
| Mean number of antibiotics | 1.1 |
| Mean number of antifungals | 2.7 |
| Mean number of treatments | 4.4 |
| Mean time to first treatment | 2.0 ± 1.6 months |
| Mean initial visual acuity (logMAR, Snellen equivalent) | 1.76 ± 0.9 ≈ 20/1200 |
| Mean final visual acuity (logMAR, Snellen equivalent) | 1.84 ± 1.2 ≈ 20/1400 |
The clinical settings included cataract surgery (9/16, 56%), glaucoma surgery (4/16, 25%), trauma (2/16, 13%), and corneal ulcer (1/16, 6%). Excluding the 2 patients whose previous surgical dates were unknown, the mean time from the predisposing event to the initial onset of symptoms and treatment was 2.0 ± 1.6 months ( Figure 1 ).
The fungal isolates were Candida (4/16, 25%, including C. albicans [2], C. parapsilosis [2]), Fusarium (3/16, 19%), Acremonium strictum (2/16, 13%), Curvularia (2/16, 13%), Paecilomyces lilacinus (2/16, 13%), Aspergillus nigricans (1/16, 6%), Phialophora richardsiae (1/16, 6%), and Helicomyces (1/16, 6%). Fourteen patients (88%) were initially treated for presumed bacterial endophthalmitis with intravitreal injections of vancomycin and ceftazidime; 2 patients (13%) received concurrent intravitreal steroids before the cultures grew fungus. Only 2 patients (13%) who presented with fungal infiltrates in the anterior chamber and vitreous cavity received intravitreal antifungal therapy as part of their initial therapy. A mean of 4.4 ± 2.8 antifungal injections and 1.1 ± 0.6 antibacterial injections were administered.
Upon identification of the fungal isolates, 5 patients (31%) received intravitreal antifungal injections while 9 (56%) had PPV with intravitreal antifungals. One patient resolved his infection without intravitreal antifungal therapy ( Figure 2 ), undergoing a pars plana vitrectomy (PPV) with removal of the intraocular lens (IOL) and capsular bag and receiving oral voriconazole. The initial intravitreal antifungal therapies were amphotericin B (12/16, 75%), voriconazole (4/16, 25%), and miconazole (3/16, 19%). Twelve patients (75%) were placed on systemic antifungal treatment in order to augment local therapy; the medications included oral voriconazole (3/16, 19%), oral diflucan (3/16, 19%), oral ketoconazole (2/16, 13%), oral natamycin (2/16, 13%), intravenous voriconazole (1/16, 6%), and oral fluconazole (1/16, 6%).
All patients underwent a PPV at least once during the course of their treatment. The 9 patients (53%) whose IOLs and capsular bags were removed had significantly better final visual acuities than those who did not (1.1 ± 1.4 logMAR, ≈20/260, vs 2.5 ± 0.8 logMAR, light perception, P = .011). Of the 7 patients whose IOLs and capsular bags were not removed, 4 proceeded to enucleations or alcohol ablation.
The fungal sensitivities were available in 9 patients (56%). The minimum inhibitory concentrations (MIC) for all isolates ranged from 0.06 to 2.1μg/mL for voriconazole, from 0.2 to 8 μg/mL for amphotericin B, and from 0.25 to 6μg/mL for fluconazole. Molds had higher median MICs compared to yeast. The sensitivities are summarized in Table 2 .
| Fungal Isolate (Number of Isolates = 9) | Amphotericin B (μg/mL) | Voriconazole (μg/mL) | Fluconazole (μg/mL) |
|---|---|---|---|
| Acremonium (1) | 1 | 0.5 | 0.5 |
| Aspergillus (1) | 8 | 0.25 | 64 |
| Candida (1) | 0.06 | 0.02 | 0.25 |
| Curvularia (1) | 0.5 | 0.06 | 4 |
| Dematicus (1) | 0.2 | – | – |
| Fusarium (2) | 1.0 | 2.1 | 64 |
| Paecilomyces (1) | 16 | 0.25 | 64 |
| Phialophora (1) | 0.5 | – | – |
a Included are 3 eyes whose specimens were previously re-cultured to determine the antifungal sensitivities.
In the current study, the mean visual acuity on presentation was 1.76 ± 0.9 logMAR (Snellen equivalent ≈20/1200), which was similar to the mean vision at the last examination (1.84 ± 1.2 logMAR, ≈20/1400, P = .83). Two patients (13%) achieved BCVAs of 20/40 or better at the last examination, and half achieved ≥20/400 (8/16). The final visual acuity was similar in mold and yeast ( P = .37). The preinfection BCVA was known in 10 patients; 4 (40%) achieved a final vision within 1 line of their pre-endophthalmitis BCVA. Patients who underwent initial vitrectomies had worse final visions than those who had initial vitreous tap-and-injects (2.2 ± 1.0 logMAR, ≈20/3000, vs 1.1 ± 1.0 logMAR, ≈ 20/220, P = .040). The 2 patients who received early intravitreal steroids had worse initial and final BCVAs than patients who did not (final BCVA: no light perception vs 1.6 ± 1.1 logMAR, ≈20/800, respectively, P = .0028). There was no difference in the final BCVAs based on the type of antifungal therapy ( P = .31), early antifungal therapy ( P = .14), or the adjunctive use of systemic antifungal therapy ( P = .63). Table 3 summarizes the mean BCVAs at the initial and last follow-up examinations.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
