A 90-year old Caucasian man came to our department in September 2017 for a routine check-up because of persistent untreated conjunctivitis in his right eye. The patient presented with hypertension treated with amlodipine and furosemide. The ocular history was positive for bilateral primary open angle glaucoma treated with Timolol 0.5%. The patient underwent ectropion surgery five years earlier (right lower lid lateral tarsal strip), bilateral cataract surgery with IOL implantation, and a later YAG-laser capsulotomy. On his left eye, an aponeurotic ptosis was present.
The clinical examination showed a best-corrected distance visual acuity (BCDVA) of 0.7 logMAR (20/100) in the right eye. Intraocular pressure was 20 mmHg in his right eye and 15 mmHg in the left eye. A slit-lamp examination showed conjunctival hyperemia and inferior brownish non-granulomatous endothelial precipitates. The corneal fluorescein staining was negative. The posterior segment did not show any pathological findings.
A viral etiology was suspected and, given that PCR analysis is not available in our clinic, an empirical treatment was started with Ganciclovir gel 1.5 mg/g, 3 times a day. Fourteen days later, the patient presented with a worsening of the clinical condition: BCDVA was hand movement, and he had a corneal paracentral ulcer (1 mm in diameter) with fluorescein staining. The ulceration was surrounded by a mild, localized corneal edema. The Tyndall effect was positive and the precipitates were more numerous.
A corneal scraping was then performed, and the clinical manifestation was highly suspected to be a fungal keratitis. Therapy was modified with 2 mg/ml topical voriconazole 1% 4/die and moxifloxacin 0.3% 6/die.
The cultural test on Sabouraud agar gave a diagnosis of
Anterior segment photography and corresponding AS-OCT image of the preoperative condition. It shows a round 4 × 4 mm corneal ulcer reaching the deep stroma, associated with massive melting surrounding the edema and perikeratic injection.
The literature has reported promising results with CXL as adjunct therapy for fungal keratitis. Because the clinical condition became worse with antifungal therapy, and since we obtained
Thirty minutes before the surgery, pain medication was administered.
The CXL-WA [
The procedure was conducted in sterile surgical conditions. The patient was draped, the ocular surface was rinsed with sterile physiologic balanced salt solution, and a lid speculum was applied.
A hypo-osmolar 0.1% riboflavin solution (RICROLIN TE Sooft, Italy) was instilled for 30 minutes before irradiation, to obtain stromal swelling. The cornea was exposed to ultraviolet light A with the UV-X System (Peschke Meditrade GmbH, Huenenberg, Switzerland), which emits light at a wavelength of l370 ± 5 nm and an irradiance of 3 mW/cm2 or 5.4 J/cm2. Exposure lasted for 30 minutes, during which time riboflavin solution was applied 6 times, every 5 minutes. After surgery, the patient received levofloxacin drops (Oftaquix; Tubilux Pharma, Pomezia, Rome, Italy). Follow-up visits entailed photo documentation with and without fluorescein, AS-OCT (Cirrus, Karl Zeiss OCT, Germany), and analysis of the anterior chamber with Pentacam (Oculus Optikgerate GmbH, Wetzlar, Germany). Before treatment therapy was 2 mg/ml topical Voriconazole 1% 4/die and Moxifloxacin 0.3% 6/die. The first day after CXL-WA, therapy was modified by adding Betamethasone 0.1% + Naphazoline 0.15% + Tetracycline 1% (Alfaflor®) eye drops 3/die and Iodopovidone 0.6% eye drops 3/die, to reduce the inflammation and to prevent bacterial contaminations; moxifloxacin 0.3% was stopped.
During the postoperative follow-up, from 1 week until the second month, the patient was reevaluated and showed a progressive reduction in the margins of the corneal opacity which was documented with anterior segment photography and AS-OCT (Figure
Anterior segment photography follow-up after CXL-WA: 1 day, 1 week, 1 month, and 2 months showing initial apparent stability of the clinical condition.
One month after CXL-WA, no signs of keratitis reactivation were present: we documented a reduction in the conjunctival hyperemia and chemosis and a complete corneal reepithelialization, as well as an improvement in corneal transparency (Figure
Until the third month after treatment, the patient underwent a weekly follow-up that confirmed corneal stability. AS-OCT and Scheimpflug images showed improvement in corneal transparency, despite a significant corneal thinning (thinnest point was 150
In the fourth month after CXL-WA, the patient showed a 2 × 2 mm central descemetocele with a thinning of the cornea (232
Anterior segment photography and AS-OCT showing a central 2x2 mm descemetocele with a thinning of the cornea (232
In order to prevent a corneal perforation, we performed
PAS positive section showing the focal erosion of the epithelium with a full thickness mixed inflammatory infiltrate, including lymphocytes, histiocytes, and granulocytes. Note the presence, in this context, of numerous branching hyphae which are highly suggestive of fungal keratitis.
Fungal keratitis is caused mainly by two types of pathogens: yeast and mold. Yeast infection is often seen among patients with ocular diseases [
CXL of corneal stromal collagen fibers, induced by ultraviolet light A and riboflavin (Vitamin B2), has been postulated to have several mechanisms of action which could aid in the management of infectious keratitis. CXL increases the stiffness and enhances the mechanical and biological stability of the stroma by promoting the formation of covalent bonds among collagen molecules in the cornea; this effect minimizes the collagen lysis caused by pathogenic microorganisms [
As previously described, we decided to use the CXL-WA technique that we first introduced in 2013 [
In this case study, we also performed a manual debridement of the superficial melting prior to CXL-WA. This helped clear out part of the superficial mycotic melting as well as allowing the cross-linking effect to directly reach the deep stromal layers. Another difference from the standard protocol is the riboflavin solution used. In the presence of an acute infection and corneal ulcer, it is not possible to determine which part of the residual pachymetry is attributable to the stroma and which is attributable to melting. For this reason, a hypo-osmolar riboflavin solution is recommended, to induce corneal swelling and to avoid the risk of the treating residual stromal bed too thin. The use of photoactivated riboflavin in the treatment of melting corneal ulcers was first proposed in 2000 [
Several case reports have shown the effectiveness of CXL: in the treatment of recalcitrant bacterial keratitis, in the improvement of the clinical symptoms, in the halting of the progressive melting, and the resolution of infectious treatment-resistant keratitis [
In vitro studies are controversial: the CXL procedure alone has not been shown to be effective against inactivate fungus, although another study showed that CXL plus amphotericin improved inhibition of fungal pathogens over amphotericin alone [
The results of the available clinical trials are also controversial. A previous trial studied the effectiveness of CXL versus antimicrobial treatment alone in patients with bacterial, fungal,
Papaioannou et al. [
A small randomized clinical trial investigating cross-linking as adjuvant therapy for deep fungal ulcers suggested that CXL could increase the rate of perforation in this condition [
Another retrospective analysis was done by Erdem [
Most of these studies display the same problem: it is difficult to have a homogenous and standardized pool of cases, due to the low incidence of this condition and the different stages of presentation. Despite this, and although there is not as much evidentiary support for using CXL to treat fungal keratitis, it is already used in conjunction with antifungals by some clinicians, in the hope that it might add some benefit, given the poor prognosis for fungal ulcers [
In summary, the literature shows no clear path to follow in advanced cases and, unfortunately, to date, the most likely scenario is still to end up with penetrating keratoplasty.
In this case study, CXL was performed as an attempt to halt the progression of the infection, and we can speculate that it helped stabilizing the cornea in the short to midterm (3 months). However, there were confounding factors that make the effects of the CXL less clear. In fact, after having performed the CXL, we modified the therapy by adding Dexamethasone; steroids are known to improve the symptoms but not the long-term outcomes. Further, because of the ongoing antifungal treatment, the initial clinical improvement could not be entirely attributed to CXL.
Nevertheless, CXL did not completely eradicate the fungal infection which caused perforation 4 months after treatment. This is confirmed by the presence of fungal hyphae in the corneal tissue analyzed during penetrating keratoplasty.
For all these reasons, we believe that CXL-WA cannot be considered the definitive solution to advanced mycotic keratitis, which still maintains a poor long-term prognosis.
The authors declare that they have no conflicts of interest.
Special thanks are extended to Giulia Raimondi for her contribution in clinical collecting and analyzing data during the first three months after treatment.