To evaluate the efficacy of photoactivated chromophore for infectious keratitis (PACK-CXL) in the treatment of patients with moderate to severe infectious keratitis as adjunct therapy to the topical medication treatment.
Randomized clinical trial.
Thirty eyes from 30 patients with moderate to severe infectious keratitis were randomized to receive either standard treatment plus PACK-CXL (n = 15) or standard treatment alone (control group, n = 15). The primary outcome was the sizes of stromal infiltrates measured on slit-lamp photographs 30 days after treatment. The secondary outcomes were the sizes of epithelial defects, the complication rates, and best pinhole-corrected visual acuity (BPVA).
The median (interquartile range [IQR]) sizes of stromal infiltrates at day 30 were 5.0 mm 2 (0–23.0 mm 2 ) in the PACK-CXL group and 10.6 mm 2 (1.1–16.3 mm 2 ) in the control group (median difference 0, 95% CI −7.0 to 0, P = .66). The median (IQR) sizes of epithelial defects were 0.7 mm 2 (0–6.3 mm 2 ) and 4.6 mm 2 (0–10.2 mm 2 ) in the PACK-CXL group and control group, respectively (median difference −3.0, 95% CI −0.8 to 0, P = .41). The complication rates and BPVA after treatment were comparable between groups.
Standard treatment combined with PACK-CXL did not provide any advantageous effect over standard treatment alone in moderate to severe infectious keratitis over a 30-day period.
Infectious keratitis is a leading cause of rapid and devastating visual loss worldwide, especially in developing countries. Despite topical broad-spectrum medical therapies being used initially, infectious keratitis leading to corneal perforation or endophthalmitis is not uncommon. A new paradigm-changing treatment able to enhance microbial eradication and improve treatment outcomes with fewer side effects needs to be established.
Corneal collagen cross-linking is a procedure in which the photosensitizer riboflavin and ultraviolet A (UVA) irradiation are used. This procedure preliminarily aims to strengthen the corneal stroma, thereby improving the corneal biomechanics in ectatic corneal disorders. Soon after the acceptance of this concept, corneal cross-linking was proposed to be effective for treating infectious keratitis based on the disinfectant properties of photoactivated chromophore. The possible mechanisms include inhibition of microbial replication, intercalation of the chromophore with microbial nucleic acid, direct damage to the pathogen cell walls by reactive oxygen free radicals, increased resistance of the cross-linked cornea to enzymatic damage, and changing of the ocular surface environment. However, the clinical evidence in terms of collagen cross-linking efficacy for keratitis is still inconclusive.
In this study, we evaluated the efficacy of photoactivated chromophore for infectious keratitis (PACK-CXL) as an adjunct to medical treatment for patients with moderate to severe infectious keratitis.
The Institutional Review Board, Faculty of Medicine, Chulalongkorn University approved and monitored this randomized controlled trial, which adhered to the tenets of the Declaration of Helsinki. The trial was registered with clinicaltrials.gov ( NCT01831206 ). The sample size for the study was calculated using a superiority design formula with power of 0.8 and a 2-tailed significance level of .05 to detect 7 mm 2 difference in areas of stromal infiltration with standard deviation of 6. This provided a sample size of 15 patients per group. Written informed consent was obtained from all participants.
The participants were recruited from the Department of Ophthalmology, King Chulalongkorn Memorial Hospital, Bangkok, Thailand from March 2013 to December 2014. All patients presenting with infectious keratitis underwent ophthalmic examination including best pinhole-corrected visual acuity (BPVA), slit-lamp biomicroscopy, anterior segment photography, and posterior segment ultrasonography. The severity of keratitis was graded by slit-lamp biomicroscopy using a modification of Jones’s grading. Ulcers that were 2-6 mm in size and infiltration that involved the mid stroma but not beyond the posterior one third of the corneal stroma were graded as moderate infectious keratitis. Ulcers either involving the posterior one third of the cornea or that were more than 6 mm in size were graded as severe infectious keratitis.
Consecutive cases of patients aged older than 6 years with moderate to severe infectious keratitis were enrolled in the study. Pregnant patients or patients with a history or evidence of herpetic keratitis, parasitic keratitis, corneal perforation, autoimmune diseases, endophthalmitis, or corneal thickness less than 400 μm by ultrasound pachymetry were excluded.
After enrollment, participants were randomized to receive standard treatment with or without PACK-CXL using simple randomization. Sealed envelopes, used to conceal the randomization, were opened after enrollment of each participant. A microbiological evaluation included Gram stain, KOH preparation, and cultures; the samples were obtained by corneal scraping in all participants.
Participants randomized to standard treatment received standard medical therapy according to the patient’s history, clinical findings, and initial laboratory results. The primary medical therapy for bacterial keratitis included hourly instillation of fortified cefazolin (50 mg/mL; BIOLAB, Samutprakarn, Thailand) and fortified amikacin (20 mg/mL; Atlantic Lab, Bangkok, Thailand); the primary medical therapy for fungal keratitis included hourly topical application of amphotericin B (1.0 mg/mL; Bharat Serums and Vaccines, Maharashtra, India) and topical natamycin (50 mg/mL; Alcon, Bangkok, Thailand). In case of positive clinical response, the medications were tapered based on the judgment of 1 clinician (N.K.). However, if the ulcers progressed, the regimens were changed according to the results of the microbiological evaluation. All participants were treated as inpatients until the medications were tapered to applications of fewer than 4 times a day. No topical or systemic corticosteroids were used during the study period.
For the participants randomized to PACK-CXL, corneal collagen cross-linking with UVA and riboflavin was performed under topical anesthesia on the first day of presentation. The corneal limbus was shielded by a Merocel ring (Medtronic, Inc, Dublin, Ireland). Riboflavin (MedioCROSS [Peschke Meditrade GmbH, Germany] 0.1% riboflavin/20% dextran solution) was administered to the cornea every 2 minutes for an initial period of 30 minutes and then every 5 minutes for a further 30 minutes during UVA illumination. The epithelium was not removed as there were epithelial defects overlying the ulcers. The UV-X lamp (Peschke Meditrade GmbH, Hünenberg, Switzerland) was used to deliver UVA (365 nm with 3.0 mW/cm 2 ) for 30 minutes. After the PACK-CXL treatment, the participants received standard medical treatment as described previously.
The results of measurement of the BPVA, slit-lamp examination, and anterior segment photography with and without fluorescein staining were recorded at the initial presentation and on day 7 and day 30. The logarithm of the minimal angle of resolution visual acuity values equal to or below the counting fingers level were recorded as follows: counting fingers, 1.7; hand movements, 1.8; light perception, 1.9; and no light perception, 2.0. All anterior segment photography was masked and 1 investigator (V.S.) randomly assessed the images for the sizes of stromal infiltrates and epithelial defects using Image-Pro version 7.0 (Media Cybernetics, Rockville, Maryland, USA). The primary outcome was the size of the stromal infiltrates; the secondary outcomes were the sizes of the epithelial defects, the complication rates after treatment, and BPVA at day 30. The results were compared between the 2 groups using Mann-Whitney U test for both the sizes of the stromal infiltrates and epithelial defects; the unpaired t test was used for the complication rates and BPVA (IBM SPSS statistics, Version 22.0; IBM Corp, Armonk, New York, USA). An alpha value of 0.05 was considered significant.
Thirty participants with infectious keratitis were enrolled in this study. The ulcers in all participants involved the visual axis. The baseline characteristics were comparable in both groups ( Table 1 ). The clinical pictures suggested that 7 patients in the PACK-CXL group and 5 patients in the control group had bacterial keratitis. Eight and 10 cases, respectively, in the PACK-CXL group and control group were treated for fungal keratitis. Five of the 15 participants (33.33%) in each group had a history of contact lens wear. Seven (46.67%) and 8 (53.33%) participants in the PACK-CXL and control groups, respectively, had a history of infection with a vegetative origin. Both participants with a Pythium infection had been exposed to contaminated water. The microbiological laboratory results are shown in Table 1 .
|Parameter||PACK-CXL Group (n = 15)||Control Group (n = 15)|
|Mean age (range), y||44.60 (17–73)||53.93 (15–84)|
|Moderate to severe infectious keratitis ratio||2:13||4:11|
|Median size of epithelial defect (IQR), mm 2||31.29 (13.48–41.61)||31.11 (19.13–45.94)|
|Median size of stromal infiltration (IQR), mm 2||31.89 (28.52–62.78)||31.07 (17.93–54.88)|
|Hypopyon||9/15 (60%)||7/15 (46.67%)|
|Mean initial BPVA (logMAR)||1.75 ± 0.22||1.68 ± 0.32|
|– Bacteria||7/15 (46.67%)||5/15 (33.33%)|
|– Fungus||8/15 (53.33%)||10/15 (66.77%)|
The median sizes of stromal infiltrates in the PACK-CXL group at days 7 and 30 were 22.1 mm 2 (interquartile range [IQR], 16.0–53.6 mm 2 ) and 5.0 mm 2 (IQR, 0–23.0 mm 2 ) and those in the control group were 25.4 mm 2 (IQR, 10.3–48.6 mm 2 ) and 10.6 mm 2 (IQR, 1.1–16.3 mm 2 ), respectively. There was no significant ( P = .50 and P = .66) differences between the groups. Subgroup analysis in the bacterial and fungal keratitis samples also showed no significant differences ( Table 2 ).
|All Cases (Median, IQR; mm 2 )||Bacterial Keratitis Cases (Median, IQR; mm 2 )||Fungal Keratitis Cases (Median, IQR; mm 2 )|
|PACK-CXL Group (n = 15)||Control Group (n = 15)||Median Difference (95% CI)||P Value a||PACK-CXL Group (n = 7)||Control Group (n = 5)||Median Difference (95% CI)||P Value a||PACK-CXL Group (n = 8)||Control Group (n = 10)||Median Difference (95% CI)||P Value a|
|Day 7||22.1 (16.0 to 53.6)||25.4 (10.3 to 48.6)||4.76 (−0.74 to 6.36)||.50||18.1 (13.6 to 44.6)||21.7 (8.1 to 25.4)||10.3 (−3.6 to 17.1)||.42||30.2 (19.9 to 64.8)||40.5 (12.5 to 59.7)||6.0 (−12.4 to 10.4)||.62|
|Day 30||5.0 (0 to 23.0)||10.6 (1.1 to 16.3)||0 (−7.0 to 0)||.66||0.8 (0.3 to 19.1)||8.6 (0 to 14.9)||0.73 (−7.8 to 4.6)||.75||9.1 (0 to 27.4)||12.6 (4.4 to 23.9)||−1.1 (−16.9 to 0)||.87|
The median sizes of the epithelial defects at days 7 and 30 in the PACK-CXL group were 23.0 mm 2 (IQR, 13.5–41.6 mm 2 ) and 0.7 mm 2 (IQR, 0–6.3 mm 2 ), respectively; the median sizes of the epithelial defects in the control group were 16.9 mm 2 (IQR, 6.3–39.5 mm 2 ) and 4.6 mm 2 (IQR, 0–10.2 mm 2 ), respectively. However, there were no significant ( P = .68 and P = .41) differences between the 2 groups. Subgroup analysis regarding etiology also showed no significant differences ( Table 3 ).