Article Text

other Versions

Collagen crosslinking in the management of advanced non-resolving microbial keratitis
  1. Rohit Shetty,
  2. Harsha Nagaraja,
  3. Chaitra Jayadev,
  4. Yathish Shivanna,
  5. Thungappa Kugar
  1. Narayana Nethralaya Eye Hospital, Bangalore, Karnataka, India
  1. Correspondence to Dr Harsha Nagaraja, Narayana Nethralaya Eye Hospital, 121/C, Chord Road, 1 “R” Block, Rajajinagar, Bangalore 560 010, Karnataka, India; harshdr{at}


Aim To evaluate the efficacy and safety of corneal collagen crosslinking (CXL) in the management of culture proven microbial keratitis.

Methods 15 eyes of 15 patients of microbial keratitis were included in the study. Nine patients had bacterial keratitis and six had fungal keratitis. All patients underwent microbiological evaluation to identify the causative organism. The depth of the infiltrate was determined clinically with slit lamp and measured manually using anterior segment optical coherence tomography. Patients were treated with antibiotics/antifungals and those who did not respond to at least 2 weeks of topical medications underwent CXL as per the standard protocol. The same preoperative topical medications were continued post-CXL. All patients were followed up every third day and observed for signs of resolution of microbial keratitis.

Results Six of nine patients with bacterial keratitis and three of six patients with fungal keratitis resolved following CXL treatment. Patients with deep stromal keratitis or endothelial plaque failed to resolve. All patients had resolution of pain on the first postoperative day. There was an appearance of or increase in hypopyon in seven patients. No intraoperative or postoperative complications were noticed.

Conclusions CXL appears to be an effective procedure in the management of superficial microbial keratitis. It can be used as an adjunctive treatment in the management of non-resolving microbial keratitis.

  • Cornea
  • Infection
  • Inflammation
  • Wound Healing

Statistics from


Corneal diseases are a major cause of vision loss and blindness according to WHO; ocular trauma and corneal ulceration result in 1.5–2 million new cases of corneal blindness annually.1 Microbial keratitis is characterised by a corneal epithelial defect with underlying stromal inflammation caused by replicating microorganisms. Microbial keratitis is a leading cause of ocular morbidity and blindness worldwide.2 Microbial keratitis requires aggressive management to halt the disease process and reduce the extent of corneal scarring, which leads to loss of vision.3 ,4

It has been estimated that about 50% of the eyes have poor visual outcome if the diagnosis and initiation of appropriate antimicrobial treatment are delayed.5 Long-term and intensive topical chemotherapy has led to the emergence of multidrug-resistant bacteria that might further complicate treatment. Microbes develop resistance to antibiotics as a result of chromosomal mutation, inductive expression of latent chromosomal genes, exchange of genetic material via transformation, bacteriophage transduction or plasmid conjugation.6 The high cost of drugs, high frequency of resistance to antibiotics, and risk of corneal melting and corneal scars make the choice of new approaches very desirable. Since 2000, riboflavin has been used as a photosensitiser to inactivate pathogens in plasma, platelet and red cell products.7 ,8 In vitro experiments have proven the bactericidal effect using a 365 nm ultraviolet-A (UV-A) light photo activation of riboflavin.9 ,10 Multiple studies have indicated that this could be a new tool in the management of infectious keratitis resistant to antibiotic treatment.11 In our study, we evaluated the role of corneal collagen crosslinking (CXL) in the management of advanced non-resolving microbial keratitis.

Materials and methods

Patients with non-resolving microbial keratitis were included in the study after taking informed consent. The study was performed under the Tenets of the Declaration of Helsinki. At initial presentation, all patients underwent corneal scrapping for Gram's stain, potassium hydroxide preparation (KOH wet mount), bacterial culture (blood agar and chocolate agar) and fungal culture (Sabouraud dextrose agar). Patients were started on topical antibiotics/topical antifungals based on the Gram's stain/KOH wet mount reports which were obtained immediately after the patient presented to us. There was no delay in starting topical medications in any of the patients. All patients were prescribed topical atropine 1% eye ointment in addition to antibacterials and antifungals. They were followed up every third day and observed for signs of resolution of the corneal involvement and hypopyon. Those patients not responding to topical chemotherapy for more than 2 weeks were considered as having non-resolving microbial keratitis involvement and advised to undergo CXL.

Patients with culture proven bacterial and fungal keratitis were included in the study. Those patients with perforated corneal ulcer, endophthalmitis, viral keratitis and pregnancy were excluded.

Procedure: Topical anaesthesia was administered using 0.5% proparacaine drops. Epithelium was not removed as there was already a large epithelial defect overlying the area of keratitis in all patients. Riboflavin drops (Medio-Cross riboflavin–dextran solution, 0.1%) were instilled topically on the cornea every 2 min for a period of 30 min. The cornea was then illuminated using a Phoenix UV-A system (Peschke Meditrade GmbH, Huenenberg, Switzerland), UVA 365 nm, with an irradiance of 3 mW/cm2 and a total dose of 5.4 J/cm2 over 30 min. During the period of UV illumination, riboflavin was administered to the eye every 2 min. Patients were continued with the same topical medications that were being used prior to CXL. Topical corticosteroids or NSAIDs were not prescribed after the crosslinking procedure.


In all, 15 eyes of 15 patients including 10 men and five women with culture proven non-healing microbial keratitis were included in the study. Mean age of the patients was 51 ± 13.38 years (range 27–71 years). Based on the smear and culture reports, nine patients had bacterial keratitis and six had fungal keratitis. The most common bacteria isolated were staphylococci and aspergillus was the most common fungus. Seven patients had hypopyon at the time of presentation.

Table 1 gives the pre-CXL and post-CXL details of all patients.

Table 1

Details of patients with non-resolving microbial keratitis showing both precollagen and postcollagen crosslinking parameters

All patients who underwent CXL had resolution of pain on the first postoperative day. In four patients with hypopyon cornea ulcer, there was an increase in hypopyon by the third postoperative day. There was an appearance of hypopyon in three patients who did not have hypopyon prior to CXL (figures 14). In patients who responded to CXL (nine out of 15), the mean time for epithelial healing was 21.3±6.14 days and the mean time for resolution of corneal infiltrate was 33.44±6.2 days. Of the seven patients who either had an increase or appearance of hypopyon, three patients underwent therapeutic keratoplasty while in the other four, the resolution of hypopyon took longer than the resolution of the corneal infiltrate.

Figure 1

Preoperative non-resolving bacterial keratitis with anterior stromal infiltrate and hypopyon.

Figure 2

Post-CXL at 8 weeks showing complete resolution of corneal infiltrate with persisting hypopyon.

Figure 3

Non-resolving fungal keratitis after 2 weeks of treatment with topical antifungals.

Figure 4

Post-CXL at 7 weeks showing complete resolution of corneal infiltrate with appearance of hypopyon.

The mean preoperative score on the Wong-Baker FACES Pain Rating Scale was 8.27±0.73 and the mean postoperative score was 0.53±0.85 indicating that there was a significant reduction in pain.

No intraoperative complications were noted.


In patients undergoing therapeutic keratoplasty for non-healing microbial keratitis, approximately a third of transplanted grafts become re-infected.12 There has therefore been a constant quest to find an alternative to therapeutic keratoplasty when medical management fails.

Tsugita et al13 showed that the combination of riboflavin and UVA caused a deactivation of the RNA in tobacco mosaic virus. Martins et al opined that a combination of UVA–riboflavin has antibacterial properties in vitro against microorganisms causing microbial keratitis.9 Hence, we evaluated the effectiveness of CXL in our patients with infective keratitis. We found that in patients with superficial corneal infiltrates involving only the anterior third or less of the stroma, there was a better response than in those cases with deeper infiltrates.

According to previous studies, the treatment seems to be most effective in blocking corneal melt caused by Gram-negative bacteria (92%) followed by Gram-positive bacteria (84%), acanthamoeba (71%) and finally fungi (61%).1420 In our experience, bacterial keratitis has a better prognosis compared with fungal keratitis probably because of the fact that most of the bacterial keratitis involved only the superficial part of the corneal stroma.

We also noticed that all patients who underwent CXL had resolution of pain on the first postoperative day. The possible explanation for this could be damage to the subepithelial nerve plexus by the riboflavin–UVA combination due to ‘chemical denervation’. The reduction of corneal sensation in all patients post-CXL supports this hypothesis.

The other important finding was that there was an increase in hypopyon in four patients post-CXL and appearance of hypopyon in three patients. We attribute this increase or appearance of hypopyon possibly due to the penetration of riboflavin through the inflamed and oedematous stroma into an already inflamed anterior chamber causing further inflammation. It was also noted that the hypopyon took a longer time to resolve than the re-epithelisation and resolution of surface infiltrate.

The main drawback of the study was that it was not known if the cases of microbial keratitis would have responded to topical medications if they were continued for a longer period of time; however, CXL did accelerate healing once it was performed.

In conclusion, CXL appears to be an effective procedure in treating non-resolving microbial keratitis with superficial stromal involvement. CXL can be an effective adjunctive treatment and add to the armamentarium of treatment modalities in the management of resistant microbial keratitis. However, further studies are required to streamline the indications, protocol and safety profile of CXL in the management of microbial keratitis.


View Abstract


  • Contributors All authors have contributed in the preparation of the manuscript and the clinical study.

  • Competing interests None.

  • Ethics approval Narayana Nethralaya Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.