Fungal Keratitis
Etiology · Clinical Features · Management
Examination question · ~800 words
What is fungal keratitis? Discuss its etiology, clinical features, investigations, and management — including the role of newer antifungals and evolving diagnostic modalities.
Try to outline your answer mentally before expanding sections below.
Fungal keratitis (keratomycosis) is a vision-threatening infectious corneal ulcer caused by filamentous fungi or yeasts. It constitutes approximately 17–36% of all microbial keratitis in tropical regions such as India (up to 40–50% in referral centres), with higher prevalence than in temperate climates. The hallmark is an indolent clinical course, delayed diagnosis, and propensity for deep stromal penetration, making it one of the most challenging corneal infections to treat. Early recognition and prompt antifungal therapy are critical to preserving vision.
1.1 Classification of Causative Organisms
Causative Organisms
| Category | Common Genera / Species | Geographic Association | Typical Risk Factor |
|---|---|---|---|
| Filamentous (Hyaline) | Fusarium solani, F. oxysporum; Aspergillus flavus, A. fumigatus | Tropical, subtropical (India, Africa, SE Asia) | Vegetative trauma (plant/thorn/soil) |
| Filamentous (Dematiaceous) | Curvularia lunata; Bipolaris, Alternaria spp. | Tropical regions | Vegetative trauma; immunosuppression |
| Yeasts | Candida albicans; C. parapsilosis, C. tropicalis | Temperate climates (Europe, USA) | CL wear, ocular surface disease, steroid use, immunosuppression |
| Dimorphic fungi | Histoplasma capsulatum; Blastomyces, Coccidioides | Endemic zones only | Rare; systemic dissemination |
1.2 Predisposing Risk Factors
Risk Factors
| Risk Factor | Mechanism / Comment |
|---|---|
| Vegetative / agricultural trauma | Most common RF in India (>60%); plant material inoculates fungi directly into corneal stroma |
| Contact lens wear | Predominant RF in developed nations; Fusarium associated with ReNu MoistureLoc solution (2006 outbreak — global recall May 2006) |
| Topical corticosteroids | Suppress local immunity; suppress PMN function; promote fungal proliferation — classic trap in misdiagnosed keratitis |
| Chronic ocular surface disease | Dry eye, bullous keratopathy, neurotrophic keratitis — impairs epithelial barrier |
| Systemic immunosuppression | HIV/AIDS, DM, post-transplant — promotes Candida and Aspergillus infections |
| Prior corneal surgery | PKP, LASIK flap interface fungal keratitis — rare but catastrophic |
| Chronic topical antibiotics | Alter conjunctival flora; ecological niche for fungal overgrowth |
1.3 Pathogenesis
After inoculation, fungi elaborate proteases, phospholipases, and mycotoxins that degrade the extracellular matrix. Filamentous fungi species differ in stromal invasion pattern: Fusarium hyphae grow parallel to corneal lamellae in the anterior stroma, enabling deep penetration without early perforation signs; Aspergillus hyphae grow vertically/perpendicularly through lamellae, accounting for greater rapid deep penetration and perforation risk. Both can track through Descemet's membrane into the anterior chamber, causing endophthalmitis even without frank perforation. Candida produces biofilm on corneal surface, resisting immune clearance and antifungal penetration. The host inflammatory response (PMNs, macrophages) contributes to stromal melting through matrix metalloproteinases (MMP-9).
2.1 Symptoms
Symptoms
| Symptom | Clinical Note |
|---|---|
| Pain | Moderate to severe; less severe than bacterial keratitis — a key differentiating clue |
| Photophobia | Present but often disproportionately mild relative to ulcer size |
| Watering / discharge | Mucopurulent discharge present but less copious than bacterial |
| Blurring of vision | Progressive; corresponds to corneal opacity extent |
| Delayed onset | Symptoms typically begin 5–14 days post-trauma (range days to weeks); indolent progression over weeks |
2.2 Slit-Lamp Signs — Ulcer Characteristics
Slit-Lamp Signs
| Sign | Description | Significance |
|---|---|---|
| Dry, rough, raised edges | Heaped-up, feathery or 'bird's feather' margins | Pathognomonic of filamentous fungi |
| Satellite lesions | Small stromal infiltrates beyond main ulcer margin | Represent hyphal extension through lamellae |
| Immune ring (Wessely ring) | Grey-white stromal ring around infiltrate | Antigen-antibody precipitate; seen in approximately 1–5% of fungal keratitis cases |
| Hypopyon | Yellow-white, dense, immobile hypopyon | Often sterile initially (toxic products); becomes infected only after Descemet's perforation or spontaneous drainage |
| Perineural infiltrates | Infiltrates tracking along corneal nerves | Characteristic of Acanthamoeba but also seen in fungi |
| Endothelial plaque | White endothelial exudate over hypopyon | Indicates pre-Descemet/endothelial involvement |
| Pigmented infiltrate | Brown/black discolouration | Dematiaceous (melanin-containing) fungi — Curvularia |
| Indolent margins | Slow-spreading; irregular, serrated edges | Contrast with rapidly spreading bacterial ulcer |
2.2b Comparison: Fungal vs Bacterial vs Acanthamoeba Keratitis
Differential Comparison
| Feature | Fungal | Bacterial (Pseudomonas) | Acanthamoeba |
|---|---|---|---|
| Pain | Moderate, indolent | Severe, acute | Severe, disproportionate |
| Onset | Subacute (days–weeks) | Acute (hours–days) | Subacute |
| Ulcer margins | Feathery, raised, dry | Suppurative, wet, undermined | Ring infiltrate, 'dirty' grey |
| Satellite lesions | Present | Absent | Absent |
| Hypopyon | Dense, immobile | Fluid, mobile | Present in severe cases |
| Risk factor | Agricultural trauma | CL, trauma, surgery | CL in tap water |
| Confocal finding | Hyphae (bright filaments) | Absent specific pattern | Double-walled cysts |
2.3 Complications
- Corneal perforation (10–30%) — requires emergency surgery
- Endophthalmitis — hyphae can penetrate INTACT Descemet's membrane without frank perforation
- Descemetocele formation
- Secondary glaucoma — trabecular blockade by inflammatory cells/fungi
- Corneal vascularisation and scarring
- Phthisis bulbi in neglected/severe cases
Laboratory Investigations
| Investigation | Method / Detail | Expected Finding |
|---|---|---|
| Corneal scraping — KOH 10–20% | Potassium hydroxide wet mount; 1st-line rapid test | Hyphae (septate/aseptate), budding yeasts |
| Corneal scraping — Gram stain | Identifies yeast and bacteria simultaneously | Gram-positive Candida; branching hyphae |
| Giemsa stain | Better for cell morphology and yeast forms | Yeast cells with pseudohyphae (Candida) |
| Calcofluor white stain | Fluorescent stain binding chitin in fungal cell wall | Bright fluorescent hyphae/yeasts; highest sensitivity among bedside stains (~95–98% with KOH) |
| Sabouraud's dextrose agar | Culture medium; 22–25°C; growth typically within 3–7 days; observe up to 2–3 weeks (4–6 weeks only for dimorphic fungi) | Gold standard for speciation and antifungal sensitivity testing |
| Potato dextrose agar | Promotes sporulation for speciation | Fusarium identification |
| Blood agar / Chocolate agar | Rule out bacterial co-infection | Sterile / bacterial growth |
| Confocal microscopy (IVCM) | In vivo; non-invasive; real-time imaging | Hyperreflective hyphae in stroma; ~80–90% sensitivity (pooled ~85–88%; observer-experience-dependent) |
| PCR (18S/ITS rDNA) | High sensitivity; rapid speciation (18S = eukaryotic fungal marker; ITS = internal transcribed spacer) | Detects fungal DNA even in culture-negative cases |
| Anterior segment OCT (AS-OCT) | Depth of infiltrate; Descemet involvement | Guides surgical decision (PTK vs DALK vs PKP) |
Calcofluor white with fluorescence microscopy is the most sensitive stain for fungi. KOH mount (10–20%) is the most practical rapid bedside test. Culture on Sabouraud's remains gold standard for speciation and antifungal sensitivity testing but results are delayed (typically 3–7 days; observe up to 2–3 weeks).
IVCM has transformed non-invasive diagnosis, particularly when corneal scraping is inconclusive. Pooled sensitivity is ~85–88%, not the often-cited 90%.
PCR should use 18S/ITS rDNA targets (fungal eukaryotic markers) — NOT 16S rRNA, which is a bacterial ribosomal marker.
4.1 Antifungal Agents — Classification and Mechanism
Antifungal Agents
| Drug Class | Agents | MOA | Spectrum / Use |
|---|---|---|---|
| Polyenes | Natamycin 5% | Binds ergosterol → sequesters membrane ergosterol → blocks transport proteins; does NOT form pores (distinct from amphotericin B) | First choice for filamentous fungi (Fusarium, Aspergillus); poor stromal penetration due to large particle size — works at epithelial/superficial stromal level |
| Polyenes | Amphotericin B 0.15% | Binds ergosterol → membrane pore formation → cell lysis | Candida (yeast keratitis). Some in vitro activity vs Aspergillus but NOT first-line for filamentous fungi |
| Azoles | Voriconazole 1%; Fluconazole 0.2%; Itraconazole | Inhibits CYP51 (lanosterol 14α-demethylase) → ergosterol depletion | Voriconazole: broadest azole; good stromal penetration. Fluconazole: Candida |
| Echinocandins | Caspofungin 0.5%; Micafungin | Inhibits β-1,3-glucan synthase → disrupts fungal cell wall | 2nd/3rd line; Candida; poor corneal penetration |
| Allylamine | Terbinafine 1% | Inhibits squalene epoxidase → squalene accumulation | Dermatophytes; limited corneal data |
4.2 Treatment Protocol — First-Line Topical Therapy
- Natamycin 5% eye drops — treatment of choice for filamentous fungi (Fusarium, Aspergillus). Dosing: every 1 hour (waking hours) for first 48–72 hours, then taper based on response.
- Voriconazole 1% topical — superior corneal penetration; achieves therapeutic stromal levels. Preferred when deep stromal involvement, Aspergillus, or natamycin failure.
MUTT-I (Prajna NV et al. JAMA Ophthalmol. 2013;131(4):422–429; n=323): The pivotal RCT for filamentous fungal keratitis. Natamycin 5% was significantly superior to voriconazole 1% for Fusarium keratitis (better BCVA, smaller scar). For non-Fusarium filamentous keratitis, there was NO statistically significant difference between natamycin and voriconazole — voriconazole was NOT shown to be superior for non-Fusarium species. The conclusion: voriconazole should not be used as monotherapy in filamentous keratitis.
MUTT-II (Prajna NV et al. JAMA Ophthalmol. 2016; PMID 27787540): Oral voriconazole as adjunct to topical antifungal therapy added no benefit for severe filamentous fungal keratitis.
Note: An earlier exploratory pilot study (Prajna et al. Arch Ophthalmol. 2010;128(6):672–678; n=120) preceded MUTT-I but found no significant difference between treatments (underpowered). This is the 2010 pilot, NOT MUTT-I. MUTT-I is the 2013 JAMA Ophthalmol paper.
4.2 Systemic Therapy
Systemic Antifungal Therapy
| Situation | Drug | Dose / Route | Duration |
|---|---|---|---|
| Deep stromal / impending perforation | Voriconazole oral | 200 mg BD (IV if severe) | 4–6 weeks; guided by response |
| Candida keratitis | Fluconazole oral | 200–400 mg/day | Till healing |
| Aspergillus / refractory | Voriconazole IV | 6 mg/kg BD load → 4 mg/kg BD maintenance | Min 6 weeks |
| Intolerance / resistance | Caspofungin IV | 70 mg load → 50 mg/day | Combination with azole |
| Endophthalmitis extension | Amphotericin B IV + Intravitreal Amp B | 0.7–1 mg/kg/day IV; 5–10 mcg intravitreal | 6–8 weeks |
4.3 Adjunctive and Surgical Management
Adjunctive and Surgical Options
| Modality | Indication / Note |
|---|---|
| Cycloplegics (atropine 1%) | Prevent synechiae; relieve ciliary spasm; adjunct in all cases |
| Stop steroids | Discontinue all topical/systemic steroids immediately — steroids promote fungal growth and must be avoided throughout active infection |
| Antifungal-soaked BCL | Bandage CL soaked in voriconazole — prolonged contact time; emerging practice |
| Corneal cross-linking (PACK-CXL) | Riboflavin-UV-A generates reactive oxygen species toxic to fungi. Growing evidence from small RCTs and a meta-analysis supports PACK-CXL as adjunct for mild–moderate fungal keratitis, reducing surgical intervention rate. Evidence base is limited; not yet standard of care. |
| Intrastromal voriconazole injection | 50 mcg in 0.1 mL (concentration 0.05% = 0.5 mg/mL) injection into infiltrate; bypasses poor topical penetration; for deep stromal/resistant cases. Established therapeutic dose from Aravind/AIIMS case series. |
| PTK (Phototherapeutic Keratectomy) | Superficial infections not responding to medical therapy; removes infected epithelium and Bowman's layer |
| DALK (Deep ALK) | Deep stromal infection sparing Descemet's — tectonic and optical; preserves host endothelium; AS-OCT guides eligibility |
| PKP (Penetrating Keratoplasty) | Impending/actual perforation; failed DALK; total corneal involvement — therapeutic PKP; 'hot eye' PKP carries high graft failure risk |
| Evisceration / Enucleation | Panophthalmitis unresponsive to all therapy; last resort |
4.4 Treatment Flowchart
Management Flowchart
| Step | Action |
|---|---|
| 1. Presentation | Suspected fungal keratitis (indolent ulcer + risk factors) |
| 2. Scraping | KOH mount + Gram + Calcofluor + Sabouraud culture ± IVCM |
| 3. Empirical start | Natamycin 5% hourly + Voriconazole 1% (if deep/severe) + Atropine + STOP steroids |
| 4. Culture result | Filamentous → continue Natamycin; Candida → switch to Fluconazole/Ampho B |
| 5. Monitor at 48–72 h | Improving → taper; static → add oral voriconazole / intrastromal injection |
| 6. 7–10 days | Minimal response → consider PACK-CXL adjunct |
| 7. Impending perforation | Therapeutic PKP or DALK (based on Descemet integrity on AS-OCT) |
| 8. Total follow-up | Minimum 6–8 weeks; visual rehab (CL/PKP) after healing |
- Antimicrobial Corneal Cross-Linking (PACK-CXL): Riboflavin/UV-A generates reactive oxygen species toxic to fungi. Growing evidence from small RCTs and a 2023 meta-analysis supports PACK-CXL as adjunct for mild–moderate fungal keratitis, reducing surgical intervention rate. Evidence base is limited; 'multiple RCTs' overstates the current evidence strength.
- Intrastromal Voriconazole: Achieves targeted high stromal concentrations (50 mcg/0.1 mL). Case series from Aravind and AIIMS demonstrate efficacy for natamycin-resistant Fusarium keratitis.
- IVCM (In Vivo Confocal Microscopy): Real-time fungal imaging without scraping; monitors treatment response; distinguishes filamentous from yeast forms; role in post-PKP monitoring for recurrence.
- Metagenomic Next-Generation Sequencing (mNGS): Identifies rare fungi not detectable on culture; turnaround typically 24–48 hours (select nanopore platforms <24 hours); emerging role in culture-negative cases.
- DALK for fungal keratitis: Growing evidence that DALK (big-bubble or manual dissection) is safe and viable for deep stromal fungal keratitis sparing Descemet's, preserving endothelium and reducing graft rejection risk.
- Drug resistance: Azole-resistant Aspergillus (CYP51/TR34-L98H mutations) increasingly reported globally; combination natamycin + voriconazole being explored in resistant filamentous cases.
Fungal keratitis remains a leading cause of corneal blindness in tropical nations, demanding a high index of clinical suspicion, early laboratory confirmation, and targeted antifungal therapy guided by organism identification. Natamycin 5% remains the cornerstone for filamentous infections, while voriconazole is invaluable for deep stromal penetration and resistant cases. The MUTT-I trial (JAMA Ophthalmol 2013) provides the strongest evidence base, establishing natamycin superiority for Fusarium and showing no benefit of oral voriconazole adjunct (MUTT-II, 2016). Emerging adjuncts — PACK-CXL, intrastromal voriconazole (50 mcg/0.1 mL), IVCM-guided monitoring, and DALK — are transforming outcomes. Surgical intervention (therapeutic PKP/DALK) remains critical for refractory or perforated cases, with timing paramount to salvaging the globe.
Trap 1 — FALSE
“Natamycin penetrates the corneal stroma well”
Natamycin has POOR stromal penetration due to its large particle size. It works primarily on the epithelium and superficial stroma. This is also why its mechanism differs from amphotericin B: natamycin sequesters ergosterol and blocks membrane transport proteins without forming pores, rather than creating ion channels. Voriconazole achieves superior stromal levels.
Trap 2 — FALSE
“Hypopyon in fungal keratitis means the infection has spread to the anterior chamber”
Hypopyon in fungal keratitis is initially STERILE — it consists of inflammatory cells and toxic products, not fungi. It becomes genuinely fungal (infected) only after Descemet's membrane perforation or spontaneous drainage allows fungal elements to enter the anterior chamber.
Trap 3 — FALSE
“MUTT-I showed voriconazole is superior for non-Fusarium filamentous fungi”
MUTT-I (Prajna NV et al. JAMA Ophthalmol. 2013;131(4):422–429; n=323) showed NO statistically significant difference between natamycin and voriconazole for non-Fusarium filamentous keratitis (regression coefficient −0.02 logMAR, P=0.81). Voriconazole was NOT shown to be superior for non-Fusarium species. The correct finding: natamycin was significantly SUPERIOR for Fusarium specifically; for non-Fusarium there was no significant difference. The trial conclusion was that voriconazole should not be used as monotherapy in filamentous keratitis.
Trap 4 — FALSE
“Steroids can be used once antifungals are started to reduce inflammation”
Steroids must be avoided throughout active fungal keratitis and only considered very cautiously post-healing, if at all. Topical and systemic steroids suppress PMN function, suppress local immunity, and actively promote fungal proliferation — this is a classic diagnostic trap (misdiagnosed bacterial keratitis treated with steroids precipitating fungal overgrowth).
Trap 5 — FALSE
“Hyphae grow perpendicular to corneal lamellae, which explains deep penetration”
Growth direction is species-specific. Fusarium hyphae grow PARALLEL to corneal lamellae in the anterior stroma, enabling deep penetration without early clinical signs of perforation risk. Aspergillus hyphae grow vertically/perpendicularly through lamellae, which actually accounts for more rapid deep penetration and greater perforation risk for Aspergillus. The claim that 'perpendicular = deeper penetration' applies to Aspergillus but the assertion that this is the pattern explaining the characteristic indolent course is incorrect — that is a Fusarium feature.
Trap 6 — FALSE
“Intracameral extension of fungal keratitis requires corneal perforation”
Hyphae can penetrate INTACT Descemet's membrane, causing anterior chamber invasion and endophthalmitis WITHOUT frank clinical perforation. This is a critical clinical trap — apparent clinical integrity of the cornea does not exclude intraocular fungal spread.
Q: What is the MUTT-I trial and what are its key findings?
A: MUTT-I (Mycotic Ulcer Treatment Trial, Part I) — Prajna NV et al. JAMA Ophthalmol. 2013;131(4):422–429 — was an RCT of 323 patients in South India comparing natamycin 5% vs voriconazole 1% for filamentous fungal keratitis. Key finding: natamycin was significantly superior to voriconazole for Fusarium keratitis (better BCVA at 3 months, smaller scar). For non-Fusarium filamentous keratitis, there was NO significant difference between the two agents. Conclusion: voriconazole should not replace natamycin as monotherapy for filamentous keratitis. Note: the 2010 pilot study (Prajna et al. Arch Ophthalmol. 2010;128(6):672–678; n=120) is the underpowered exploratory precursor, NOT MUTT-I.
Q: What did MUTT-II show about oral voriconazole?
A: MUTT-II (Prajna NV et al. JAMA Ophthalmol. 2016; PMID 27787540): Oral voriconazole added as adjunct to topical antifungal therapy provided NO additional benefit over topical therapy alone for severe filamentous fungal keratitis. This has important clinical implications — routine addition of oral voriconazole is not supported by evidence.
Q: Why is natamycin the first-line agent for Fusarium keratitis despite poor stromal penetration?
A: MUTT-I demonstrated natamycin's clinical superiority for Fusarium keratitis in terms of BCVA outcomes and scar size, despite its known pharmacokinetic limitation of poor stromal penetration beyond the epithelium. Fusarium hyphae grow parallel to anterior stromal lamellae — accessible to natamycin's effective epithelial/superficial stromal concentrations. The MUTT-I evidence base definitively established natamycin as first-line for Fusarium, overriding earlier assumptions based on penetration data alone.
Q: Describe the PCR molecular markers used for fungal versus bacterial keratitis diagnostics.
A: For fungal (eukaryotic) diagnostics: 18S rRNA gene (eukaryotic small ribosomal subunit) and ITS (Internal Transcribed Spacer) regions (ITS1/ITS2/5.8S). For bacterial diagnostics: 16S rRNA gene (prokaryotic ribosomal marker). Using 16S for fungal detection is a molecular biology error — 16S is a bacterial marker. The correct PCR target for fungal keratitis is 18S/ITS rDNA or ITS/28S rDNA.
Q: What is the correct dose for intrastromal voriconazole injection and why does it matter?
A: The established therapeutic dose is 50 mcg in 0.1 mL (concentration 0.05% = 0.5 mg/mL), as used in case series from Aravind Eye Hospital, AIIMS, and published literature (Srinivasan M et al. Ophthalmology 2012). A dose of 2 mcg/0.1 mL is 25-fold below therapeutic and would be sub-therapeutic, leading to treatment failure. This is a critical clinical safety point — intrastromal injections at sub-therapeutic doses may give false reassurance while allowing progressive infection.
A 45-year-old farmer presents with a 3-week corneal ulcer after paddy field injury. IVCM shows hyperreflective filaments to 400 µm depth. AS-OCT reveals Descemet's membrane intact. Culture grows Aspergillus flavus. He has failed topical natamycin + voriconazole for 10 days. Discuss surgical decision-making: DALK vs therapeutic PKP, intraoperative conversion risk, and post-operative antifungal strategy.
Surgical Decision Framework
AS-OCT showing intact Descemet's membrane with hyphae at 400 µm depth is the key deciding factor. This supports DALK (Deep Anterior Lamellar Keratoplasty) over PKP: the host endothelium is intact and worth preserving. DALK removes all infected anterior stroma while preserving the endothelium, eliminating endothelial rejection risk and providing better long-term graft survival compared to PKP. The big-bubble technique (type 1 Dua bubble: pneumatic separation of stroma from Descemet's) is preferred. Manual dissection (Melles' technique) is the fallback if big-bubble fails.
Intraoperative Conversion Risk
Despite intact Descemet's pre-operatively, intraoperative perforation risk during DALK is 10–20% in infected corneas due to: (1) fungal proteases weakening stromal architecture; (2) big-bubble injection forces risking Descemet detachment over friable tissue. If perforation occurs, immediate conversion to PKP is necessary. The surgical team must be fully prepared for conversion from the outset with a compatible donor cornea in the sterile field. Incomplete DALK with residual infected deep stroma is not acceptable — conversion to PKP is preferable to leaving infection behind.
Post-operative Antifungal Strategy
Continue topical voriconazole 1% QID for minimum 6–8 weeks post-operatively. Add systemic voriconazole oral 200 mg BD for 6 weeks given Aspergillus (for which voriconazole is the azole of choice, IV loading if severe). Monitor with IVCM — hyperreflective elements at graft-host junction indicate recurrence. Taper antifungals only after two consecutive IVCM examinations show no hyphae and clinical signs are resolved. Avoid topical steroids until fungal clearance is confirmed; then introduce with extreme caution to prevent rejection episodes from triggering renewed fungal activity.
Why Not PKP First?
PKP in an active fungal eye ('hot eye PKP') carries significantly higher graft failure risk from: (1) ongoing inflammation reducing endothelial cell survival; (2) risk of recurrent infection in the new graft; (3) lifelong immunosuppression in a patient with Aspergillus history increasing reinfection risk. PKP is reserved for: actual Descemet's perforation detected on AS-OCT/intraoperatively, failed DALK, or total corneal involvement extending to the limbus. If DALK successfully clears the infection, PKP for optical rehabilitation can be planned later on a quiet eye — superior outcomes.
References
- Bowling B. Kanski's Clinical Ophthalmology: A Systematic Approach. 8th ed. Elsevier; 2016. ISBN: 9780702055720.
- Yanoff M, Duker JS. Ophthalmology. 5th ed. Elsevier; 2019.
- Prajna NV, Mascarenhas J, Krishnan T, et al. Comparison of natamycin and voriconazole for the treatment of fungal keratitis [pilot study]. Arch Ophthalmol. 2010;128(6):672–678. PMID 20547942.
- Prajna NV, Krishnan T, Mascarenhas J, et al. The Mycotic Ulcer Treatment Trial, Part I (MUTT-I): a randomised trial comparing natamycin vs voriconazole for filamentous fungal keratitis. JAMA Ophthalmol. 2013;131(4):422–429. PMID 23710492.
- Prajna NV, Krishnan T, Rajaraman R, et al. Effect of oral voriconazole on outcomes of fungal keratitis in the Mycotic Ulcer Treatment Trial II (MUTT-II). JAMA Ophthalmol. 2016;134(12):1365–1372. PMID 27787540.
- Thomas PA, Kaliamurthy J. Mycotic keratitis: epidemiology, diagnosis and management. Clin Microbiol Infect. 2013;19(3):210–220. PMID 23398543.
- Sharma N, Sachdev R, Jhanji V, Titiyal JS, Vajpayee RB. Therapeutic keratoplasty for microbial keratitis. Curr Opin Ophthalmol. 2010;21(4):293–300. PMID 20531191.
- Hafezi F, Randleman JB (eds). Corneal Collagen Cross-Linking. SLACK Inc.; 2013.
- American Academy of Ophthalmology. Cornea/External Disease Panel. Preferred Practice Pattern — Bacterial Keratitis. AAO; 2019.
- Gain P, Jullienne R, He Z, et al. Global Survey of Corneal Transplantation and Eye Banking. JAMA Ophthalmol. 2016;134(2):167–173. PMID 26633035.
- Garg P, Rao GN. Corneal ulcer: Diagnosis and management. Community Eye Health. 1999;12(30):21–23. PMC1706003.