Methods

A retrospective chart review was conducted for all patients with orthokeratology-associated infectious keratitis admitted to the Hong Kong Eye Hospital between January 2003 and December 2013. Cases were identified using the diagnostic codes “Corneal disorder due to contact lens” and “Keratitis” from the Clinical Data Analysis and Reporting System of the Hong Kong Hospital Authority. Exclusion criteria included non-orthokeratology cases and cases treated on an outpatient basis. Medical records were traced, and clinical information was retrieved for analysis. Patients were diagnosed to have infectious keratitis clinically under slit-lamp examination by the presence of an epithelial defect, stromal infiltrate, or specific signs such as perineuritis. Corneal scrapings were performed in all cases for microbiological investigations, including Gram stain. Microbiological samples were inoculated onto blood agar, chocolate agar, Sabouraud’s dextrose agar, and non-nutrient agar culture plates. Contact lenses and storage solutions were also obtained for microbiological cultures. All patients received intensive empirical broad-spectrum antimicrobial or amoebicidal therapy. The treatment was modified according to culture and sensitivity results as well as the clinical response after admission.

The study was approved by the Institutional Review Board of the hospital and adhered to the tenets of the Declaration of Helsinki. Statistical analysis was performed using PASW software version 18.0 (SPSS/IBM, Inc, Chicago, Illinois, USA). P values of .05 or less were considered to be statistically significant.

Results

Overall, 23 patients (16 female, 7 male) were included in this study ( Table 1 ). The mean age of patients was 15.0 ± 4.2 years (range: 9–23 years). All patients had unilateral ocular involvement (14 right eyes, 9 left eyes). Patients were using overnight orthokeratology for an average of 2.7 ± 2.8 years (range: 3 months–10 years) before the onset of infection. There were no other associated risk factors. All patients claimed to have good contact lens hygiene and did not recall any prior ocular injury. Moreover, none of them suffered from ocular surface diseases or systemic illness. Fourteen patients had received previous treatment from other ophthalmologists in the form of amoebicidal eye drops, topical fluoroquinolones, aminoglycosides, and antiherpetic eye ointment. The mean duration from onset of disease to hospital admission was 16.5 ± 30.2 days (range: 1–122 days). All patients presented with painful red eyes and reduced visual acuity. The best-corrected logMAR visual acuity at presentation was 0.62 ± 0.51 (20/83 Snellen). Clinical features and microbiological profile of all patients are summarized in Table 2 . A central or paracentral corneal infiltrate was noted in 14 eyes (60.9%) with an average size of 2.7 ± 1.7 mm (range: 0.5–6.0 mm). There was an associated hypopyon in 3 eyes (13.0%). Corneal perineuritis suggestive of Acanthamoeba keratitis was noted in 12 eyes (52.2%).

Fifteen eyes (65.2%) had a positive microbiological culture obtained from corneal scrapings ( Table 2 ). One pathogen was identified in 12 cases, while 3 cases had polymicrobial infection. The most commonly isolated organism was Pseudomonas aeruginosa (n = 6), followed by coagulase-negative Staphylococcus (CNS) (n = 5) and Acanthamoeba (n = 3). Staphylococcus aureus , Serratia marcescens , Micrococcus luteus , and Flavobacterium species each accounted for a single case. Orthokeratology lenses or lens storage solution were available for microbiological investigations in all cases. There were 5 cases of P aeruginosa and 5 cases of Acanthamoeba identified from contact lenses or solution. When the results of corneal scraping, contact lenses, and storage media were combined, the culture-positive rate was 78.2% (n = 18), and the most common pathogens were P aeruginosa (n = 8), Acanthamoeba (n = 6), and CNS (n = 5). No fungal organisms were isolated in any of the cases, although 1 case responded well to combined antibacterial and antifungal treatments.

At the time of presentation, empirical treatment was started in the form of hourly levofloxacin 0.5% or moxifloxacin 0.5% eye drops and once-daily ofloxacin 0.3% ointment. For ulcers involving the visual axis and stromal infiltrates larger than 2 mm, fortified vancomycin (50 mg/mL) and fortified tobramycin or gentamicin (14 mg/mL) were administered hourly. Polyhexamethylene biguanide 0.02% and propamidine isethionate 0.1% were started in cases suspected to have Acanthamoeba keratitis. Overall, 12 eyes, which presented with corneal perineuritis, were treated empirically with amoebicidal drugs. The mean duration of hospitalization was 15.3 ± 9.8 days (range: 5–31 days).

All patients responded to medical treatment. The mean duration from disease onset to remission was 31.9 ± 34.9 days (range: 6–131 days). Residual central or paracentral corneal scarring was observed in all cases. The final best-corrected logMAR visual acuity was 0.15 ± 0.20 (20/28 Snellen), which was significantly better than the visual acuity at presentation (Wilcoxon signed rank test, P < .001).

Considering the 2 most common pathogens in our series, the duration from symptom onset to hospitalization was 2.9 ± 2.6 days (range: 1–8 days) and 15.7 ± 12.4 days (range: 2–31 days) in cases with P aeruginosa and Acanthamoeba , respectively (Mann-Whitney U test, P = .008). The disease duration in cases with P aeruginosa and Acanthamoeba was 13.0 ± 7.7 days (range: 6–28 days) and 35.3 ± 16.1 days (range: 19–62 days), respectively (Mann-Whitney U test, P = .005). The amoebicidal drugs were continued in all 12 patients with perineuritis for up to 6 months after the initial diagnosis.

Overall, during the study period 393 patients with infectious keratitis and 121 patients with contact lens–related keratitis were admitted to the hospital for management. The proportion of orthokeratology-related keratitis cases was 5.9% and 19%, respectively, during this time period ( Table 1 ). Interestingly, of the 32 cases of contact lens–related Acanthamoeba keratitis over the 10-year period, 12 (37.5%) were attributable to orthokeratology lenses ( Table 1 ).