Fungal endophthalmitis is a vision-threatening infection most commonly found in patients with predisposing risk factors such as recent hospitalization, surgery, immunosuppression, organ transplantation, intravenous drug use, or indwelling catheters.1 Endophthalmitis is thought to occur via hematogenous spread of a disseminated infection (most commonly Candida, followed by Aspergillus1), with initial seeding of the retinal or choroidal vessels followed by tissue destruction and intraocular extension.2 It usually presents as yellow-white creamy chorioretinal lesions in the setting of blurred vision and floaters, but the initial diagnosis is often challenging and frequently missed. Treatment involves systemic and intravitreal antifungal therapy, with escalation to pars plana vitrectomy (PPV) if clinically indicated. An estimated one-third of patients develop severe visual loss (visual acuity of 20/200 or worse).3 Visual outcome depends on location of the lesion, virulence of the infectious organism, and development of potential complications such as choroidal neovascularization and retinal detachment.1
Opportunistic infections are common in immunocompromised children, particularly in those with hematologic malignancies. We present a case of presumed endogenous fungal endophthalmitis in a pediatric patient with a history of high-risk B-cell acute lymphoblastic leukemia with recovered immune function and pan-negative cultures. We highlight the role of the ophthalmologist in guiding therapy for patients with ocular and central nervous system disease in the absence of definitive laboratory testing and report the utility of a vitrectomy in a patient refractory to initial systemic and intravitreal antifungal therapy.
A 15-year-old boy with a history of high-risk B-cell acute lymphoblastic leukemia and asthma presented with a 5-day history of visual changes in his left eye. He was diagnosed as having acute lymphoblastic leukemia at 12 years of age and completed his final cycle of maintenance chemotherapy per the Childrens' Oncology Group Phase III clinical trial AALL1131 at 1 month prior to presentation. Visual symptoms were described as “colorful pinwheels” lasting several minutes followed by progressive blurring of vision. He had an associated headache that was similar to his chronic headaches, but otherwise felt well and was afebrile. He reported no recent travel, animal/tick bites, exposures, or preceding illness.
On initial examination, his visual acuity was 20/20 in the right eye and 20/300 eccentrically in the left eye. Pupils were equally reactive with no afferent pupillary defect. Visual fields were full to confrontation in the right eye and showed a superonasal defect in the left eye. Slit-lamp and dilated fundus examination was normal in the right eye but revealed 1+ white cells in the anterior chamber, a bi-lobed white chorioretinal infiltrate along the inferior arcade with mild overlying vitreous opacities, and subretinal fluid in the macula in the left eye (Figure 1A). There was no conjunctival or scleral injection, hypopyon, optic nerve edema, or retinal hemorrhages in either eye.
Ocular and central nervous system findings at presentation. (A) Fundus photograph of the left eye reveals a large, elevated, bi-lobed white chorioretinal perivascular infiltrate with overlying vitritis and associated subretinal fluid in the macula. (B) Axial T2 magnetic resonance imaging scan of the brain at the level of the corpus collosum showing a small ring-enhancing lesion in the left occipital lobe (arrow).
Laboratory work-up revealed an unremarkable complete blood count (white blood cell count = 6.3 × 109/L, no blasts, normal absolute neutrophil count), basic metabolic panel, C-reactive protein, and a slightly elevated erythrocyte sedimentation rate (28 mm/hr; normal: 0 to 20 mm/hr). Magnetic resonance imaging (MRI) of the brain and orbits revealed small ring-enhancing lesions in the left occipital lobe (Figure 1B) and posterior right parietal lobe with moderate surrounding vasogenic edema, and nonspecific scleral thickening in the left eye.
At this point, the etiology of the chorioretinal infiltrates was unclear. Although a leukemic infiltrate could not be definitively ruled out (despite lack of a blast crisis), the appearance of the infiltrates was most suggestive of a fungal rather than a malignant lesion based on the presence of anterior chamber cells and vitritis, lesion morphology, location, and lack of hemorrhages. In addition, the neuroradiologic finding of a ring-enhancing lesion was thought to be most consistent with an infectious rather than a leukemic infiltrate. The patient was prescribed intravenous ciprofloxacin (400 mg twice daily), voriconazole (6 mg/kg twice daily), and acyclovir (10 mg/kg every 8 hours) while an infectious work-up was pursued. Topical prednisolone four times a day and cyclopentolate twice a day were also initiated in the left eye to manage intraocular inflammation. Lumbar puncture, bone marrow biopsy, beta-D-glucan, cerebrospinal fluid, blood cultures, and computed tomography of the sinuses, chest, abdomen, and pelvis were all unremarkable. Cerebrospinal fluid and serum studies were also negative for herpes simplex virus, cytomegalovirus, toxoplasmosis, and cryptococcal antigen. His coverage was narrowed to intravenous voriconazole and he was monitored closely for signs of clinical improvement.
One week later, despite stable visual acuity and improving anterior chamber inflammation, there was an interval increase in vitreous opacities and a new area of vitreous breakthrough overlying the inferior chorioretinal lesion. An intravitreal injection of voriconazole (100 µg/0.1 mL) was given and he was transitioned from intravenous to oral voriconazole (100 mg twice daily). Intravitreal amphotericin B was also considered at this point, but based on significant improvement of the ring-enhancing parietal lesion on repeat MRI (not shown), it was thought that the causative agent was most likely fungal and sensitive to voriconazole. Despite these interventions, subsequent examinations revealed unchanged visual acuity and increasing vitritis. His systemic voriconazole levels were found to be subtherapeutic and dosing was increased accordingly. Intravenous amphotericin B therapy was initiated but promptly discontinued due to severe side effects.
Due to a deteriorating clinical course, a diagnostic and therapeutic PPV with vitreous biopsy, core vitrectomy, and intravitreal injections of voriconazole (50 µg/0.1 mL) and amphotericin B (5 µg/0.1 mL) was performed. Vitreous infiltrates overlying the deeper chorioretinal lesion were sampled intraoperatively and sent for hematopathology and microbiology testing. Hematoxylin–eosin staining of the vitreous sample showed numerous neutrophils, lymphocytes, and a large amount of degenerating cells and debris, with no leukemic cells or fungal elements identified (Figure 2). Vitreous cytomegalovirus, Toxoplasma, fungal cultures, and fungal ribosomal polymerase chain reaction (PCR) test results were all negative.
Hematoxylin–eosin stain (original magnification ×40) of vitreous biopsy of the left eye demonstrating lymphocytes, granulocytes, degenerating cells, and debris, with no morphologic evidence of acute leukemia or fungal elements.
The patient's clinical course steadily improved, as evidenced by decreasing vitritis, consolidation and regression of the chorioretinal infiltrates, resolved subretinal fluid with residual macular lipid (Figure 3A), and complete resolution of the ring-enhancing brain lesions (Figure 3B). He completed a 3-month course of oral voriconazole (150 mg twice daily) and continued to do well without any systemic recurrence. His ocular lesion regressed and appeared fibrotic without exudation, and his visual acuity improved to 20/20 in the left eye.
Ocular and central nervous system findings after systemic/intravitreal antifungal therapy and vitrectomy. (A) Fundus photograph of the left eye 6 weeks after vitrectomy showing interval resolution of vitritis, consolidation of the chorioretinal lesion, and resolution of subretinal fluid with residual macular lipid. (B) Axial T2 magnetic resonance imaging scan of the brain at the level of the corpus collosum 2 weeks after presentation demonstrating a significant decrease in size, enhancement, and vasogenic edema associated with the previously documented left occipital lobe lesion (arrow; compare to Figure 1B).
We present a case of presumed fungal chorioretinitis that progressed to endophthalmitis in a patient with leukemia receiving systemic and intravitreal antifungal therapy, eventually requiring a PPV with repeat intravitreal antifungal administration. This case presents several clinical challenges, particularly in differentiating a leukemic from an infectious retinal infiltrate in an otherwise healthy-appearing patient with no evidence of relapsed disease and determining the timing of surgical treatment options.
Although the diagnosis of fungal endophthalmitis is often presumptive based on clinical examination, infectious infiltrates can appear similar to leukemic infiltrates, especially in patients who cannot mount a robust immune response and may have minimal vitritis, if any.4 Therefore, a tissue biopsy may have been helpful in directing early treatment in our patient. However, neither the brain lesions nor submacular infiltrates were amenable to biopsy due to the risks of permanent neurologic damage or vision loss, respectively. In addition, the amount of tissue that could be obtained from an aqueous and/or vitreous biopsy was likely insufficient to yield a definitive histopathologic diagnosis due to minimal intraocular inflammation at the time of presentation. Thus, a systemic work-up becomes crucial to differentiate malignant from infectious infiltrates, with leukemic infiltrates being more common in patients in blast crisis and opportunistic infections more likely in patients in remission, especially in those with a history of bone marrow transplantation, steroid use, or eye pain on presentation.4 In our case, there were examination findings (intraocular inflammation and lesion morphology) and compelling systemic evidence (ring-enhancing brain lesions and lack of blast crisis) that suggested a fungal rather than leukemic etiology.
In the absence of a definitive diagnosis, treatment decisions should be guided by clinical judgment and risk analysis. A leukemic infiltrate in the eye warrants prompt initiation of orbital radiation to preserve vision, but this may result in permanent visually significant sequelae, including corneal decompensation, neovascular glaucoma, cataract formation, radiation retinopathy, and radiation-induced optic neuropathy.5 For these reasons, radiation therapy is generally reserved for patients with a definitive biopsy result or optic nerve enhancement on orbital MRI suggestive of leukemic infiltration. Conversely, broad-spectrum antibiotic and antifungal agents are generally well tolerated. Therefore, in patients with chorioretinal infiltrates it is important to determine the most likely clinical scenario and consider treating infectious causes first. Because infections are more likely to have a rapid progression compared to leukemic infiltrates, this approach expedites treatment while avoiding the permanent effects of radiation injury in patients less likely to have relapsed disease. Our patient was initially treated with antibiotic and antifungal agents based on clinical suspicion for an infectious etiology. His improvement (despite no treatment for a leukemic recurrence) provided further support for an infectious cause and helped direct subsequent antifungal therapy. Had he instead received irradiation, his disease would have been redefined as relapsed and the appropriate therapy that ultimately preserved his vision may have been delayed.
There are no established guidelines for the indication and timing of vitrectomy in the treatment of endogenous fungal endophthalmitis, but the general consensus is that a trial of systemic antifungal therapy may be used in choroiditis or early peripheral endophthalmitis, whereas early vitrectomy is preferred in cases with significant vitritis or macular involvement.2,3,6 Vitrectomy was first described for fungal endophthalmitis in the 1970s and is almost always used in conjunction with intravitreal and systemic antifungal medications with good ocular penetration.6 Besides providing diagnostic material to identify the causative agent, vitrectomy decreases the fungal load, clears inflammatory mediators, exposes loculated microabscesses, increases intraocular diffusion of systemic antifungals, and reduces the risk of retinal detachment from vitreoretinal traction.3,7 Based on the results of a recent retrospective case series, Birnbaum and Gupta8 proposed that early vitrectomy may lead to a prompt diagnosis and improved visual acuity in patients with endogenous fungal endophthalmitis.
Positive vitreous culture rates in patients with endogenous fungal endophthalmitis are variable (38% to 75%), with yields likely lower in those patients already receiving antifungal therapy.9,10 PCR detection of fungal ribosomal DNA and beta-D-glucan testing was also attempted in this case but was negative. These findings, along with the absence of any fungal structures seen on histologic analysis of the vitreous debris, could indicate that the infection was already adequately treated with systemic and intravitreal voriconazole and that only degenerated, nonviable organisms remained at the time of vitrectomy. Additionally, vitreous sample components could have acted as inhibitors to PCR analysis, thereby leading to a false-negative result.11 Almeida et al.12 recently reported decreased culture positivity rates with small-gauge (23- and 25-gauge) compared to 20-gauge vitrectomy for endophthalmitis, leading to speculation that the higher vitreous cutting rates and vacuum of small-gauge instrumentation may lead to increased shearing forces and decreased microbiological yield. Although 25-gauge instrumentation was used in our case, in vitro data have not shown a statistical difference in the yield of bacterial culture with respect to instrument gauge or cutting rate.12 Thus, it is reasonable to speculate that the negative cultures in our patient were more likely the result of prior antifungal therapy and/or inadequate sample volume rather than surgical technique and instrumentation.
Overall, our patient had an excellent response to antifungal therapy and subsequent vitrectomy. Although his infection was atypical due to lack of fever, normal white blood cell count, negative cultures, and immunocompetent status, he did have significant immunosuppression throughout his aggressive chemotherapy course up until 6 weeks prior to presentation. Of note, his absolute lymphocyte count reached a nadir of 210 cells/µL 6 months prior to presentation, which only partially recovered to 540 cells/µL (normal: 970 to 3,260 cells/µL) 2 months prior to presentation. Because T cells are critical to the immune response against fungal antigens, it is possible that he had delayed clinical manifestations of a disseminated fungal infection that developed during this time. There have been reports of delayed and suboptimal immune reconstitution after therapy for acute lymphoblastic leukemia, particularly in the setting of high-risk disease or in patients with transfusion-related immunomodulation.13
Chorioretinal infiltrates in patients with leukemia can be diagnostically challenging and require a systemic work-up and multidisciplinary approach to direct appropriate therapy. In the absence of a definitive diagnosis, infectious causes should be considered first and intervention should be guided by clinical judgment. Our case illustrates that endogenous fungal endophthalmitis can present as a delayed opportunistic infection in a seemingly immunocompetent patient and can be successfully treated with a combination of systemic and intravitreal antifungal therapy and vitrectomy.
- Sridhar J, Flynn HW, Kuriyan AE, Miller D, Albini T. Endogenous fungal endophthalmitis: risk factors, clinical features, and treatment outcomes in mold and yeast infections. J Ophthalmic Inflamm Infect. 2013;3:60. doi:10.1186/1869-5760-3-60 [CrossRef]
- Ghoraba HH, Ellakwa AF, Elgemai EM, Mansour HO, Heikal MA. Results of pars plana vitrectomy for the management of endogenous fungal endophthalmitis after urinary tract procedures. Retin Cases Brief Rep. 2017;11:171–174. doi:10.1097/ICB.0000000000000321 [CrossRef]
- Sallam A, Taylor SR, Khan A, et al. Factors determining visual outcome in endogenous Candida endophthalmitis. Retina. 2012;32:1129–1134. doi:10.1097/IAE.0b013e31822d3a34 [CrossRef]
- Gordon KB, Rugo HS, Duncan JL, et al. Ocular manifestations of leukemia: leukemic infiltration versus infectious process. Ophthalmology. 2001;108:2293–2300. doi:10.1016/S0161-6420(01)00817-X [CrossRef]
- Jeganathan VS, Wirth A, MacManus MP. Ocular risks from orbital and periorbital radiation therapy: a critical review. Int J Radiat Oncol Biol Phys. 2011;79:650–659. doi:10.1016/j.ijrobp.2010.09.056 [CrossRef]
- Smiddy WE. Treatment outcomes of endogenous fungal endophthalmitis. Curr Opin Ophthalmol. 1998;9:66–70. doi:10.1097/00055735-199806000-00012 [CrossRef]
- Lavine JA, Mititelu M. Multimodal imaging of refractory Candida chorioretinitis progressing to endogenous endophthalmitis. J Ophthalmic Inflamm Infect. 2015;5:54. doi:10.1186/s12348-015-0054-z [CrossRef]
- Birnbaum FA, Gupta G. The role of early vitrectomy in the treatment of fungal endogenous endophthalmitis. Retin Cases Brief Rep. 2016;10:232–235. doi:10.1097/ICB.0000000000000238 [CrossRef]
- Sallam A, Lynn W, McCluskey P, Lightman S. Endogenous Candida endophthalmitis. Expert Rev Anti Infect Ther. 2006;4:675–685. doi:10.1586/14787188.8.131.525 [CrossRef]
- William A, Spitzer MS, Deuter C, et al. Outcomes of primary transconjunctival 23-gauge vitrectomy in the diagnosis and treatment of presumed endogenous fungal endophthalmitis. Ocul Immunol Inflamm. 2016:1–7.
- Wiedbrauk DL, Werner JC, Drevon AM. Inhibition of PCR by aqueous and vitreous fluids. J Clin Microbiol. 1995;33:2643–2646.
- Almeida DR, Chin EK, Shah SS, et al. Comparison of microbiology and visual outcomes of patients undergoing small-gauge and 20-gauge vitrectomy for endophthalmitis. Clin Ophthalmol. 2016;10:167–172.
- Kah TA, Yong KC, Rahman RA. Disseminated fusariosis and endogenous fungal endophthalmitis in acute lymphoblastic leukemia following platelet transfusion possibly due to transfusion-related immunomodulation. BMC Ophthalmol. 2011;11:30. doi:10.1186/1471-2415-11-30 [CrossRef]