Original Article

Evaluation of corneal collagen cross-linking as anadditional therapy in mycotic keratitisRasik B Vajpayee MS FRANZCO,1,2 Shah N Shafi MD,1 Prafulla K Maharana MD,1,3 Namrata Sharma MD1 andVishal Jhanji MD2,4

1Dr Rajendra Prasad Centre for Ophthalmic Sciences, All India Institute of Medical Sciences, New Delhi, 3LV Prasad Eye Institute,Bhubaneswar, Orissa, India; 2Centre for Eye Research Australia, University of Melbourne, Melbourne, Victoria, Australia; and4Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong SAR, China

ABSTRACT

Background: To report the treatment outcomes ofmycotic keratitis with collagen cross-linking.

Design: Retrospective study.

Participants: Patients with smear-positive moderatemycotic keratitis.

Methods: A retrospective case-file analysis was per-formed to identify cases of moderate mycotic kerati-tis treated with and without additional collagencross-linking, in addition to intensive topical anti-fungal therapy. Patients in which collagen cross-linking was performed on the day of presentation(group 1) were compared with patients whoreceived medical treatment alone in the form of 5%natamycin eye drops (group 2).

Main Outcome Measures: The primary outcomemeasure was the time taken for resolution of infec-tion.

Results: Overall, 41 cases were included for analysis(group 1, 20 cases; group 2, 21 cases). Mean age of thepatients was comparable in both groups (46.5 ± 17.01vs. 41.2 ± 20.7 years; P = 0.36). Average infiltrate sizewas 16.35 ± 6.8 mm2 in group 1 and 17.09 ± 7.4 mm2

in group 2 (P = 0.83). Overall, Aspergillus was themost commonly isolated organism (n = 4 group 1;

n = 6 group 2). Resolution of infection was observedin 18 cases (90%) in group 1 and 18 (85.71%) cases ingroup 2. The average healing time was 30.85 ± 26.6days in group 1, while it was 31.28 ± 19.97 days ingroup 2 (P = 0.94). Final best-corrected visual acuityin group 1 was 1.13 ± 0.55 and 1.25 ± 0.46 in group 2(P = 0.46). A tectonic keratoplasty was performed intwo cases in group 1 and three cases in group 2(P = 1.00).

Conclusions: In our study, additional collagen cross-linking treatment did not have any advantage overmedical management in cases with moderatemycotic keratitis.

Key words: collagen cross-linking, fungal keratitis,mycotic keratitis, outcomes, treatment.

INTRODUCTION

Mycotic keratitis continues to be a major challengefor ophthalmologists.1,2 It is relatively more commonin warmer climates, such as India, compared withthe temperate regions of the United States.2 Thereported proportion of mycotic keratitis ranges from8% of the infectious ulcers in Western countries3 toas high as 50% in India.2 In general, mycotic keratitisis more difficult to treat as compared with bacterialkeratitis4 since the antifungal medications have poorpenetration especially in the presence of an intact

■ Correspondence: Professor Rasik B Vajpayee, Centre for Eye Research Australia, University of Melbourne, 32 Gisborne Street, East Melbourne, Vic.

3002, Australia. Email: rasikv@unimelb.edu.au

Received 11 February 2014; accepted 6 July 2014.

Competing/conflicts of interest: No stated conflict of interest.

Funding sources: No stated funding sources.

Meeting presentation: Presented as a poster in the annual meeting of the Association for Research in Vision and Ophthalmology, 2013.

bs_bs_banner

Clinical and Experimental Ophthalmology 2014; ••: ••–•• doi: 10.1111/ceo.12399

© 2014 Royal Australian and New Zealand College of Ophthalmologists

epithelium.5 New treatment approaches, such asintrastromal targeted drug delivery, have beenreported to improve the treatment outcomes espe-cially in cases with recalcitrant mycotic keratitis.6

More recently, collagen cross-linking (CXL) has beenproposed as a treatment option for mycotic keratitis.Variable clinical success has been reported with itsuse as an adjuvant therapy.7–10 We report the out-comes of use of CXL in cases with moderate mycotickeratitis in a tertiary care hospital in North India.

METHODS

All cases with smear-positive mycotic keratitis pre-senting to the Cornea Clinic of Dr Rajendra PrasadCentre for Ophthalmic Sciences, New Delhi,between January 2011 and August 2012 were ana-lysed retrospectively. An institutional review boardapproval was obtained for retrospective analysis ofthe data. The study adhered to the tenets of the Dec-laration of Helsinki. All cases with moderate fungalkeratitis11,12 with the largest infiltrate diameter<6 mm and or involving <60% of the corneal thick-ness as seen on slit-lamp examination wereincluded. Cases with diabetes mellitus, immunosup-pression or collagen vascular disorders, impendingperforation, and presence of endothelial plaque orhypopyon were excluded. The following informa-tion was noted from the medical records: demo-graphic profile, risk factors for keratitis (such ashistory of trauma, contact lens use, steroid use,surgery), slit-lamp biomicroscopic findings (size,depth, location and area of the ulcer; anteriorchamber reaction; presence and height of hypopyon),microbiological results and final outcome at the endof 3 months.

All identified cases were divided into two groups:group 1 in which CXL was performed on the day ofpresentation in addition to the commencement ofmedical therapy (natamycin 5%; Natamate, SUNCompanies, Mumbai, India); and group 2, in whichthe patients received medical treatment alone. As perthe protocol followed at our centre, the patients wereadmitted for management. All patients receivednatamycin 5% eye drops hourly for 2 days and every2 h thereafter with subsequent tapering dependingupon the clinical response. In addition, patientswere offered CXL treatment as an additional treat-ment after explaining the details of the treatmentbased on the available evidence. The patients wereasked to make a decision if an offer for CXL wasmade.

CXL was performed on the day of presentationusing a standard surgical technique. Briefly, topicalanaesthesia with 0.5% proparacaine hydrochloride(Paracaine; Sunways, Mumbai, India) was usedbefore the surgery. The corneal epithelium was

debrided over the area of infiltrate. One drop of iso-tonic riboflavin phosphate 0.1%(10 mg of riboflavin-5-phosphate in 10 mL of dextran-T-500 20% solution,Medio-cross GmbH, Neudorf, Germany) was appliedevery 3 min for 30 min before irradiation and every3 min during irradiation. Ultraviolet A (UVA) radia-tion (365 nm with the desired irradiance of 3 mW/cm2; UV-X, IROC, Zurich, Switzerland) was appliedat 5 cm from the cornea for 30 min. Medical manage-ment was continued after the CXL.

The primary outcome measure was time to reso-lution of infection. Secondary outcomes includedbest-corrected visual acuity (BCVA, measured on thelogMAR scale) and the rate of corneal perforation at3 months. In addition, the size and depth of theresidual scar, vascularization, and adverse effects, ifany, were recorded, and results were comparedbetween the two groups.

Statistical analysis

The data were recorded on an Excel sheet (Microsoft,Redmond, WA, USA). The baseline characteristicswere compared between the two groups to detectany imbalance in the two groups. The categoricalvariables were summarized by frequency (%), andthese were compared between the two groups usingthe chi-square test or Fisher exact test as appropri-ate. The quantitative values were summarized asmean ± standard deviation, and non-parameterictests were used to compare difference in mean valuesbetween the two groups. For the primary outcomevariable effect size, the difference in mean values ofbest spectacle corrected visual acuity (BSCVA) on thelogMAR scale and its 95% confidence interval werecomputed. The criteria for the results were based ona null value not included in the 95% confidenceinterval. Stata statistical software version 11.0 (StataCorp, College Station, TX, USA) was used for thedata analysis.

RESULTS

Overall, 41 cases were included (group 1, 20 cases;group 2, 21 cases). Mean age of the patients was46.55 ± 17.01 years in group 1 and 41.23 ± 20.7 yearsin group 2 (P = 0.36) (Table 1). The most commonassociated risk factor was trauma with vegetativematter, reported in six cases in group 1 and five casesin group 2. Thirteen cases in group 1 and 14 cases ingroup 2 did not have an identifiable risk factor.Overall, Aspergillus was the most commonly isolatedorganism (n = 4 group 1; n = 6 group 2) (Table 2).

Average size of the infiltrate was 16.35 ± 6.8 mm2

in group 1 and 17.09 ± 7.4 mm2 in group 2 (P = 0.83).LogMAR BCVA at presentation was 1.38 ± 0.46 ingroup 1 and 1.49 ± 0.43 in group 2 (P = 0.37).

2 Vajpayee et al.

© 2014 Royal Australian and New Zealand College of Ophthalmologists

Resolution of infection was observed in 18/20cases (90%) in group 1 and 18/21 (85.71%) cases ingroup 2. The average healing time was 30.85 ± 26.6days in group 1 and 31.28 ± 19.97 days in group 2(P = 0.94). A tectonic keratoplasty was performed forperforated corneal ulcer in two cases in group 1 andthree cases in group 2. At the end of 3 months, mostof the cases had superficial vascularization (85%group 1, 90.4% group 2), while deep vascularizationwas seen in few cases (15% group 1, 9.52% group 2).There was no statistically significant differencebetween the two groups in any of the postoperativeparameters (Table 3). Final BCVA was 1.13 ± 0.55and 1.25 ± 0.46 in group 1 and group 2, respectively,(P = 0.46) (Table 3).

DISCUSSION

The indications of CXL have expanded in the recentyears for cases with corneal ectasia, bullouskeratopathy and infectious keratitis. Although wellestablished for treatment of progressive keratoconusand corneal ectasia, the role of CXL in infectiouskeratitis is still under evaluation. The antimicrobialeffects of CXL have been attributed to several mecha-nisms. First, it may have a direct cytotoxic effect onthe organisms by inactivation of ribonucleicacids.13,14 Second, it increases the tissue resistance toenzymatic digestion by the pathogens,15 therebyenhancing the corneal resistance against microbialdegradation and corneal melting.16 Third, cellapoptosis after CXL may also affect the inflammatorycells, thereby reducing the inflammatory responseand symptoms associated with infectious keratitis.17

The role of CXL in the management of infectiouskeratitis has been explored in the recent years.7–10

Compared with bacterial keratitis, the managementof fungal keratitis poses a challenge mainly due to adelayed diagnosis and insufficient treatment options.Even with the availability of newer antifungal drugswith high bioavailability like voriconazole,18–20

fungal corneal ulcers can be more difficult to treatthan bacterial corneal ulcers. A recent surveyrevealed that 80% of corneal specialists believed thatexisting treatments for fungal ulcers were only mod-erately effective.21 There is a constant need to expand

Table 1. Comparison of baseline parameters between two groups of patients with moderate mycotic keratitis

Parameters Group 1 Group 2 P-value(medical treatment and collagen cross-linking) (medical treatment)

Mean age (years) 46.55 ± 17.01 41.23 ± 20.7 0.36Males/females 14/7 11/10 0.53Mean infiltrate size (mm2) 16.35 ± 6.8 17.09 ± 7.4 0.83BCVA at presentation 1.38 ± 0.46 1.49 ± 0.43 0.37

BCVA, best-corrected visual acuity in logMAR.

Table 2. Preoperative and postoperative parameters ofpatients with moderate mycotic keratitis

Number Age(years)

Microbiologicalculture

Treatment Healing time(days)

1 72 Fusarium CXL + N 602 50 Aspergillus CXL + N 143 42 CXL + N 304 26 CXL + N 145 45 Aspergillus CXL + N 426 41 CXL + N 427 60 CXL + N 358 28 CXL + N 499 45 CXL + N 30

10 36 Fusarium CXL + N Cornealperforation

11 65 CXL + N 9012 12 CXL + N 2113 60 Aspergillus CXL + N 3114 55 CXL + N 2615 39 CXL + N 1516 67 CXL + N 2917 44 CXL + N 4018 45 Aspergillus CXL + N 3219 60 Fusarium CXL + N Corneal

perforation20 39 CXL + N 1721 10 Curvularia N 6022 35 N Corneal

perforation23 20 Fusarium N 2124 40 Aspergillus N 4925 8 N 4926 48 N Corneal

perforation27 75 Aspergillus N 3028 39 N 2129 62 N 9030 52 N 3031 30 Fusarium N 3032 70 Aspergillus N 4233 30 N 3034 20 N 3035 56 N 3036 55 Aspergillus N Corneal

perforation37 46 N 2838 38 Aspergillus N 3939 24 N 2540 48 N 1841 60 Aspergillus N 35

CXL, collagen cross-linking; N, topical natamycin.

Collagen cross-linking in mycotic keratitis 3

© 2014 Royal Australian and New Zealand College of Ophthalmologists

the armamentarium of treatment options available tocombat mycotic keratitis. Encouraged by the reportsof success of cross-linking in the management ofinfectious keratitis,7–10 we started offering thismodality of treatment to patients with moderatemycotic keratitis, in addition to medical manage-ment. In the current study, we compared the treat-ment outcomes in cases with moderate fungalkeratitis with and without adjunctive CXL. Resolu-tion of infection was seen in 85.7% of the cases thatreceived medical treatment, whereas 90% of thecases healed with combined medical treatment andCXL. The addition of CXL to the treatment protocoldid not affect the time to resolution or final visualoutcome in our study.

In an earlier study, Li et al. performed CXL in eightcases of microbiologically proven fungal keratitisthat did not respond to conventional treatment.8

Resolution of infection and epithelial healing wasachieved in all the cases after CXL. In addition, noneof the cases required corneal transplantation.10 Inanother recent study, Price et al. performed CXL inseven cases with fungal keratitis with variable infil-trate diameters. The authors reported an overall poorresponse to treatment without intensive adjuncttopical or intracorneal injections of antifungal medi-cations.9 In our study, we performed CXL as anadjunct treatment to evaluate its benefit. We did notfind any added advantage of CXL over intensivemedical treatment in cases with mycotic keratitis.There was no significant difference in the durationof healing, perforation rate or final BCVA. Sincethe effect of CXL is limited to the anterior 400 μmof stroma only,22 we deliberately chose cases thatinvolved anterior 60% of the corneal stroma.However, the determination of ulcer depth wasbased on clinical examination in our study; it is pos-sible that the mycotic infection was deeper micro-scopically.

In vitro studies have found that riboflavin/UVAis effective against Staphylococcus aureus, Staphylococcusepidermidis, Pseudomonas aeruginosa, Methicillin-resistant Staphylococcus aureus, multidrug-resistantPseudomonas aeruginosa and drug-resistant Streptococ-cus pneumoniae, but is ineffective against Candida

albicans.23 Sauer et al. reported that CXL alone hasno in vitro effect, but previous treatment withamphotericin B allowed riboflavin/UV-A to be effec-tive against Candida albicans, Fusarium sp andAspergillus fumigatus.24

It is noteworthy that medical treatment in theform of natamycin eye drops was successful for themanagement of majority of the cases with moderatefungal keratitis in our study. A randomized clinicaltrial showed that natamycin treatment was associ-ated with significantly better clinical and micro-biological outcomes than voriconazole treatmentfor smear-positive filamentous fungal keratitis.25

Natamycin is the only topical antifungal approvedby the US Food and Drug Administration fortopical ophthalmic use. It is considered to be thefirst line of therapy against mycotic keratitis incountries where it is available. In our study, com-bination of CXL and natamycin did not seem tooffer any significant benefit over the use ofnatamycin alone. The average cost of CXL isUS$1500 approximately. Therefore, it is practicallydifficult to use CXL as an adjuvant unless its ben-efits are sharply defined.

The main drawback of our study is its retrospec-tive design. Also, we did not debride the cornealepithelium in cases receiving only medical treat-ment. Nevertheless, our study did not show anyadvantage of using CXL as an adjuvant in cases ofmoderate mycotic keratitis. Further randomizedclinical trials would be required in the future toassess the role of CXL in the treatment of infectiouskeratitis.

REFERENCES

1. Whitcher JP, Srinivasan M, Upadhyay MP. Cornealblindness: a global perspective. Bull World Health Organ2001; 79: 214–21.

2. Srinivasan M, Gonzales CA, George C et al. Epidemi-ology and aetiological diagnosis of corneal ulcerationin Madurai, South India. Br J Ophthalmol 1997; 81:965–71.

3. Varaprasathan G, Miller K, Lietman T et al. Trends inthe etiology of infectious corneal ulcers at the F. I.Proctor Foundation. Cornea 2004; 23: 360–4.

Table 3. Comparison of postoperative parameters between two groups of patients with moderate mycotic keratitis

Parameters Group1 Group 2 P-value(medical treatment and collagen cross-linking) (medical treatment)

Final BCVA 1.13 ± 0.55 1.25 ± 0.46 0.46Scar size (mm2) 14.39 ± 6.20 15.16 ± 6.88 0.71Corneal perforation 2 (10%) 3 (14.2%) 1.00Time to healing (days) 30.85 ± 26.6 31.28 ± 19.97 0.94Superficial vascularization 17 (85%) 19 (90.47%) 0.66Deep vascularization 3 (15%) 2 (9.52%) 0.66

BCVA, best-corrected visual acuity in logMAR.

4 Vajpayee et al.

© 2014 Royal Australian and New Zealand College of Ophthalmologists

4. Srinivasan M. Fungal keratitis. Curr Opin Ophthalmol2004; 15: 321–7.

5. Tuli SS. Fungal keratitis. Clin Ophthalmol 2011; 5:275–9.

6. Sharma N, Chacko J, Velpandian T et al. Comparativeevaluation of topical versus intrastromal voriconazoleas an adjunct to natamycin in recalcitrant fungal kera-titis. Ophthalmology 2013; 120: 677–81.

7. Anwar HM, El-Danasoury AM, Hashem AN. Cornealcollagen crosslinking in the treatment of infectiouskeratitis. Clin Ophthalmol 2011; 5: 1277–80.

8. Li Z, Jhanji V, Tao X et al. Riboflavin/ultravoiletlight-mediated crosslinking for fungal keratitis. Br JOphthalmol 2013; 97: 669–71.

9. Price MO, Tenkman LR, Schrier A et al. Photoactivatedriboflavin treatment of infectious keratitis using colla-gen cross-linking technology. J Refract Surg 2012; 28:706–13.

10. Makdoumi K, Mortensen J, Crafoord S. Infectiouskeratitis treated with corneal crosslinking. Cornea2010; 29: 1353–8.

11. Jones DB. Decision-making in the management ofmicrobial keratitis. Ophthalmology 1981; 88: 814–20.

12. Lin HC, Lin JL, Lin-Tan DT, Ma HK, Chen HC. Earlykeratectomy in the treatment of moderate Fusariumkeratitis. PLoS ONE 2012; 7: e42126. doi: 10.1371/journal.pone.0042126.

13. Tsugita A, Okada Y, Uchara K. Photosensitized inacti-vation of ribonucleic acids in the presence of ribofla-vin. Biochem Biophys Acta 1965; 103: 360–3.

14. Goodrich RP. The use of riboflavin for inactivation ofpathogens in blood products. Vox Sang 2000; 78: 211–5.

15. Spoerl E, Wollensak G, Seiler T. Increased resistance ofcrosslinked cornea against enzymatic digestion. CurrEye Res 2004; 29: 35–40.

16. Hovakimyan M, Guthoff RF, Stachs O. Collagencross-linking: current status and future directions.J Ophthalmol 2012; doi: 10.1155/2012/406850.

17. Wollensak G, Spoerl E, Wilsch M et al. Keratocyteapoptosis after corneal collagen cross-linking usingriboflavin/UVA treatment. Cornea 2004; 23: 43–9.

18. Ramakrishnan T, Constantinou M, Jhanji V, VajpayeeRB. Factors affecting treatment outcomes withvoriconazole in cases with fungal keratitis. Cornea2013; 32: 445–9.

19. Jhanji V, Sharma N, Mannan R, Titiyal JS, VajpayeeRB. Management of tunnel fungal infection withvoriconazole. J Cataract Refract Surg 2007; 33: 915–7.

20. Jhanji V, Yohendran J, Constantinou M, Sheorey H,Vajpayee RB. Scedosporium scleritis or keratitis orboth: case series. Eye Contact Lens 2009; 35: 312–5.

21. Loh AR, Hong K, Lee S et al. Practice patterns in themanagement of fungal corneal ulcers. Cornea 2009; 28:856–9.

22. Spoerl E, Mrochen M, Sliney D et al. Safety of UVA-riboflavin cross-linking of the cornea. Cornea 2007; 26:385–9.

23. Martins SA, Combs JC, Noguera G et al. Antimicrobialefficacy of riboflavin/UV-A combination (365 nm) invitro for bacterial and fungal isolates: a potential newtreatment for infectious keratitis. Invest Ophthalmol VisSci 2008; 49: 3402–8.

24. Sauer A, Letscher-Bru V, Speeg-Schatz C et al. In vitroefficacy of antifungal treatment using riboflavin/UV-A(365 nm) combination and amphotericin B. InvestOphthalmol Vis Sci 2010; 51: 3950–3.

25. Prajna NV, Krishnan T, Mascarenhas J et al. Themycotic ulcer treatment trial: a randomized trialcomparing natamycin versus voriconazole. JAMAOphthalmol 2013; 131: 422–9.

Collagen cross-linking in mycotic keratitis 5

© 2014 Royal Australian and New Zealand College of Ophthalmologists