Clinical Features and Visual Outcome of Infectious Keratitis Associated with Orthokeratology Lens in Korean Pediatric Patients
Article information
Abstract
Purpose
To investigate the clinical features and visual outcome of infectious keratitis associated with orthokeratology (Ortho-K) lens in Korean pediatric patients.
Methods
We retrospectively reviewed medical records of patients diagnosed with Ortho-K lens-related infectious keratitis from June 2005 to April 2020 at a tertiary referral hospital. Patients’ demographics, clinical features, microbiological evaluation, and treatment methods were assessed, and factors related to final visual outcomes were analyzed.
Results
The study included 26 eyes from 26 patients (19 female and 7 male patients; mean age, 11.9 years), with an average Ortho-K lens wear duration of 33.7 ± 21.2 months. The highest number of cases occurred in summer (11 of 26 cases, 42.3%). Central or paracentral corneal lesions were observed in 25 cases (96.2%), with a mean corneal epithelial defect size of 5.13 mm2. Pseudomonas aeruginosa was the most commonly isolated organism (n = 5), followed by Serratia marcescens (n = 4). All patients responded to medical treatment without needing surgical intervention. 72% of cases achieved favorable visual outcomes (Snellen best-corrected visual acuity [BCVA] >6/12), while 8% experienced severe visual impairment (Snellen BCVA ≤6/60) due to residual central corneal opacities. Multivariable analysis showed that non-summer seasons (p = 0.043), duration from symptom onset to presentation (p = 0.040), and corneal epithelial defect size (p = 0.002) were significantly associated with final logarithm of the minimum angle of resolution BCVA. Failed autorefraction at presentation due to an Ortho-K-related infectious keratitis lesion was a significant predictor of poor final visual outcome (Snellen BCVA ≤6/12; odds ratio, 38.995; p = 0.030).
Conclusions
Ortho-K lens-related infectious keratitis can lead to permanent corneal opacities and potentially devastating visual outcomes in children. Delayed time to presentation, large corneal lesions, failure of autorefraction, and non-summer seasons were associated with poorer outcomes. Proper education and early detection would be key to safe use of orthokeratology lenses in pediatric patients.
Orthokeratology (Ortho-K) utilizes specially designed reverse-geometry gas-permeable contact lenses to correct mild myopia [1]. The use of highly oxygen-permeable material has enhanced the safety of overnight wearing, eliminating the need for glasses or contact lenses during the day. South Korea has one of the highest prevalence rates of myopia, with a 64.6% prevalence among children under 18 years old [2]. Recent studies have shown that Ortho-K lens, either alone or in combination with low-dose atropine, are effective in myopia control, leading to increased use of Ortho-K lens among myopic children [2–5].
Ortho-K lens works by flattening the central cornea, redistributing the corneal epithelium to correct myopia [6]. However, the mechanical stress exerted by the base curve of Ortho-K lens on the central cornea can lead to side effects such as corneal staining and infectious keratitis [1]. As the degree of myopia increases, the mechanical stress on the central cornea also increases [7]. This mechanical stress can cause corneal erosion, making the cornea more susceptible to infection during overnight wear [8]. Infectious keratitis is the most serious ocular complication associated with Ortho-K lens wear, potentially threatening vision if not treated promptly. The incidence of Ortho-K–associated infectious keratitis in children under the age of 18 has been reported to be between 5.3 and 13.9 per 10,000 patient-years [9,10]. Pseudomonas aeruginosa keratitis and Acanthamoeba keratitis are known to be the primary causes of serious infectious keratitis in both contact lens and Ortho-K lens wearers [11–15]. To our knowledge, while a few cases of Acanthamoeba keratitis associated with Ortho-K lens have been reported in Korean pediatric patients, there has been no study on the clinical features of Ortho-K lens-related infectious keratitis [16].
Therefore, we aim to investigate the clinical features and visual outcome of infectious keratitis associated with Ortho-K lens in Korean pediatric patients.
Materials and Methods
Ethics statement
This study was approved by the Institutional Review Board of Seoul National University Bundang Hospital (No. B-2407-915-104), with a waiver of informed consent due to its minimal risk to subjects. The study was conducted in accordance with the Declarations of Helsinki
Study design
We retrospectively reviewed the medical records of patients diagnosed with Ortho-K lens-related infectious keratitis at Seoul National University Bundang Hospital (Seongnam, Korea), a referral-based tertiary hospital, from June 2005 to April 2020. Only pediatric patients aged 18 years or younger were included. For the included patients, we investigated demographic features including age, sex, initial symptom, season of the first visit, and best-corrected visual acuity (BCVA); characteristic findings of the initial corneal lesion (involved zone and direction); microbiological culture results; treatment methods; final visual outcomes, including final BCVA and refractive errors; and factors associated with final visual outcome. To describe the characteristic findings of the initial corneal lesion, we classified the involved zone based on the following criteria: The central zone of the cornea extends from the center to 2.0 mm in diameter. The paracentral zone surrounds this central zone and extends from 2.0 to 5.0 mm in diameter. The peripheral zone surrounds the paracentral zone and extends from 5.0 to 9.0 mm in diameter. Additionally, to describe the location of the main corneal lesion, we divided the cornea into four quadrants based on its center: superior, inferior, nasal, and temporal.
Cultures were attempted from corneal scraping, lens, and lens case whenever possible. Since patients had often already used topical or systemic antibiotics before referral to our tertiary hospital, the causative organism was specified if it was isolated from any one of the three sources: corneal scraping, lens, or lens case. The source of isolation is detailed in Table 1. If Acanthamoeba is strongly suspected clinically, the culture specimen is sent to the Tropical Medicine Laboratory at Seoul National University College of Medicine (Seoul, Korea), where it is microscopically examined for the presence of trophozoites and cysts dispersed in saline, without any special staining. The specimen is then inoculated into both Escherichia coli-coated solid medium and liquid medium with peptone-yeast-glucose, to assess the growth of Acanthamoeba.
Clinical features of orthokeratology lens-related infectious keratitis in Korean pediatric patients
Severe visual impairment was defined as Snellen BCVA of 6/60 or less, moderate visual impairment as greater than 6/60 but 6/18 or less, and mild visual impairment as greater than 6/18 but 6/12 or less [17]. A poor final visual outcome was defined as mild to severe visual impairment, while a favorable visual outcome was defined as Snellen BCVA greater than 6/12.
Based on the month of the first hospital visit, spring was defined as March to May, summer as June to August, fall as September to November, and winter as December to February.
Main outcome
The primary outcomes include clinical features such as initial BCVA, corneal lesion characteristics (size and location), isolated microorganisms, seasonal occurrence, treatment methods, and final BCVA. The secondary outcome was the factors associated with the final visual outcome in the affected eye.
Statistical analysis
We performed statistical analyses using IBM SPSS ver. 19.0 (IBM Corp). The significance of the difference in visual acuity and refractive errors before and after treatment was tested using the Wilcoxon signed rank test. To test the significance between the two independent groups, we first assessed normality and then performed either an independent samples t-test or a Mann-Whitney U-test. To assess the significance of seasonal occurrence, we performed a chi-square goodness-of-fit test with equal expected proportions for each season. Linear regression analysis was conducted to investigate factors associated with the final logarithm of the minimum angle of resolution (logMAR) BCVA. Additionally, logistic regression analysis was used to investigate prognostic factors associated with poor final visual outcome (Snellen BCVA ≤6/12). Variables with correlation coefficients below 0.6, tolerance above 0.4, and variance inflation factor below 3, identified through collinearity analysis, were included in the multivariable regression analysis. Statistical significance was set at p < 0.05.
Results
Demographic features
A total of 26 eyes from 26 patients (19 female and 7 male patients) who were referred to Seoul National University Bundang Hospital from June 2005 to April 2020 for infectious keratitis associated with Ortho-K lens use were included (Table 1). Demographic and clinical features are summarized in Table 2. The mean age at presentation was 11.9 ± 3.0 years (range, 8 to 18 years), with 73.1% being 13 years old or younger (Fig. 1). All patients had unilateral ocular involvement (14 right eyes, 12 left eyes). The mean duration of Ortho-K lens wear before the onset of infection was 33.7 ± 21.2 months (range, 9 to 72 months). In four cases (15.0%), tap water was used for lens cleaning. For all included patients, there was no coexisting ocular surface diseases, systemic diseases, or prior ocular injuries. The mean duration from initial symptom onset to presentation was 3.38 ± 4.77 days (range, 1 to 22 days). The most common initial symptom was eye pain, reported in 22 cases (84.6%), followed by conjunctival injection in 17 (65.4%), tearing in four (5.4%), and conjunctival discharge, decreased visual acuity, and glare each in two cases (7.7%). Seasonal occurrence showed that the highest number of cases presented in summer, accounting for 11 (42.3%), followed by spring with six (23.1%), winter with five (19.2%), and fall with four cases (15.4%); however, this difference was not statistically significant (chi-square test, p = 0.243) (Fig. 2). Mean initial logMAR BCVA was 1.25 ± 1.16 (range, 0.00 to 3.00), with 15 of 25 patients (60.0%) experiencing severely to moderately decreased visual impairment (Snellen BCVA ≤6/18). The mean initial intraocular pressure was 15.0 ± 3.3 mmHg (range, 9 to 20 mmHg). At the time of the first visit, the proportion of failed autorefraction tests due to corneal lesions like corneal epithelial defect or infiltration was 45% (9 of 20 cases). The mean initial spherical equivalent for the remaining patients who successfully measured refractive errors was −2.90 ± 1.37 (range, −5.25 to −0.88).
Demographic and clinical characteristics of Korean pediatric patients referred to a tertiary hospital for infectious keratitis following Ortho-K lens wear
Age distribution of orthokeratology (Ortho-K) lens-related infectious keratiti in Korean pediatric patients. The proportion of patients aged 13 years and younger accounts for 73.1%, demonstrating a decreasing trend in the frequency of Ortho-K lens-related infectious keratitis with increasing age.
Seasonal distribution of orthokeratology (Ortho-K) lens-related infectious keratitis in Korean pediatric patients. The Y-axis represents the frequency of infectious keratitis associated with Ortho-K lens as a percentage of total patients.
Clinical characteristics of the initial corneal lesion
Fig. 3 illustrates the involved zone and direction of corneal lesions based on Tables 1 and 2. Corneal lesions, such as infiltration, were located in the central or paracentral zone in 25 of 26 cases (96.2%), with a mean epithelial defect size of 5.13 ± 6.08 mm2 (range, 0.04 to 21.62 mm2). The main corneal lesion direction was most commonly in the temporal (14 of 25 cases, 56.0%), followed by the nasal (7 of 25 cases, 28.0%), and the superior and inferior quadrants each with 2 of 25 cases (8.0%). An associated hypopyon was present in six eyes (23.0%). Corneal perineuritis, suggestive of Acanthamoeba keratitis, was noted in five eyes (19.2%).
Distribution of corneal lesions at initial presentation in Korean pediatric patients with orthokeratology lens-related infectious keratitis (based on Table 2). (A) Involved corneal zone of lesions. The central corneal zone was involved in 57.7% of cases, the paracentral zone in 38.5%, and the peripheral zone in 3.8%. (B) Direction of corneal main lesion. The main corneal lesion was most commonly located in the temporal quadrant (56.0%), followed by the nasal quadrant (28.0%), the superior quadrant (8.0%), and the inferior quadrant (8.0%).
Microbiological culture
The prevalence of etiological agents identified in microbiological cultures is illustrated in Fig. 4. Positive microbiological cultures were obtained from corneal scraping, lens, and/or lens case in 12 of 21 cases (57.1%) where cultures were performed. The most commonly isolated organism was P. aeruginosa (n = 5, 23.8%), with one patient showing suspected coinfection with Acinetobacter haemolyticus, followed by Serratia marcescens (n = 4, 19.0%), Achromobacter denitrificans (n = 1, 4.8%), Acinetobacter baumannii (n = 1, 4.8%), and Pandoraea species (n = 1, 4.8%). In one patient, S. marcescens was isolated in bilateral lens cultures, despite not being identified in the corneal scraping culture. In two patients, no microorganisms were isolated from the corneal scraping culture, but S. marcescens or A. baumannii were identified in the lens case, respectively. Fungal organisms and Acanthamoeba were not isolated.
Prevalence of etiological agents in microbiological cultures from Korean pediatric patients with orthokeratology lens-related infectious keratitis. In this study, among the 26 included patients, 21 underwent culture tests, with 12 of them testing positive for causative microorganisms, resulting in a culture positivity rate of 57.1%. The most common etiological agent was Pseudomonas aeruginosa, identified in a total of five patients, with one patient involving coinfection with Acinetobacter haemolyticus. Additionally, Serratia marcescens was detected in four patients, while Achromobacter denitrificans, Acinetobacter baumannii, and Pandoraea species were each found in one patient.
Treatments
The treatment methods for each patient are summarized in Table 1. Out of a total of 26 patients, the topical medication used prior to the first presentation was unknown for 11 patients. Among the 15 patients whose previous topical medications were known, 11 (73.3%) were using topical fluoroquinolones such as moxifloxacin, levofloxacin, gatifloxacin, and ofloxacin; one patient (6.6%) was using fortified ceftazidime and tobramycin; two patients (13.3%) were using both topical fluoroquinolones and tobramycin; and one patient (6.6%) was using tobramycin alone. At the time of presentation, empirical treatment was primarily initiated with fourth-generation fluoroquinolones such as moxifloxacin 0.5% and gatifloxacin 0.5%, adjusting the dosage frequency from four times a day to every 1 hour according to severity, in accordance with the randomized controlled trial study by Shah et al. [18]. If there is a positive response, the medication is continued; if the response is slow, we prefer to add fortified antibiotics. However, in cases of severe initial presentation or various other circumstances where fortified antibiotics are deemed more appropriate, we use fortified antibiotics as the primary treatment. Although Acanthamoeba was not isolated in the culture, polyhexamethylene biguanide (PHMB) 0.02% was used in five eyes which showing corneal perineuritis, suspecting Acanthamoeba keratitis. In four cases, antifungal agents were added due to poor response to antibacterial agents and/or PHMB treatment, resulting in an effective response, although no fungal organisms were isolated in any cases. All patients responded to medical treatment without needing surgical intervention. The mean duration from presentation to improvement of corneal lesions was 3.3 ± 6.2 days (range, 1 to 32 days). The mean duration from presentation to the start of antibiotic tapering was 6.2 ± 6.6 days (range, 1 to 25 days). The mean treatment duration was 50.1 ± 36.3 days (range, 9 to 115 days).
Final visual outcomes and refractive errors
The mean follow-up period was 6.5 ± 11.3 months (range, 0.1 to 43.5 months). The mean logMAR BCVA significantly improved from 1.25 ± 1.16 at presentation to 0.24 ± 0.40 at the last visit (independent samples t-test, p < 0.001). Favorable visual outcomes were achieved in 72% of cases (Snellen BCVA >6/12), while 8% experienced severe visual impairment (Snellen BCVA ≤6/60) due to residual central corneal opacities. In 17 of 20 cases (85%), residual corneal opacity was observed in the central or paracentral zone. The mean final refractive error was −3.72 ± 3.21 diopters (spherical equivalent; range, −11.88 to +2.75 diopters), while the mean initial refractive error was −2.90 ± 1.37 diopters (spherical equivalent; −5.25 to −0.88 diopters; Wilcoxon signed rank test, p = 0.116). There was no significant difference in astigmatism by refractometer between the last visit and the presentation in the affected eye (Wilcoxon signed rank test, p = 0.680). Additionally, we compared the changes in spherical and cylindrical diopters between the affected eye and the fellow eye at the last visit, assuming symmetric corneal curvature in both eyes, for the 13 patients who had their refractive errors measured at the last visit. For these patients, nine patients had central and four had paracentral corneal opacities at the last visit. The spherical diopter showed a hyperopic change in the affected eye compared to the fellow eye, though this difference was not statistically significant (−2.44 ± 2.73 diopter sphere [Dsph] [range, −6.50 to +3.50 Dsph] vs. −3.18 ± 2.48 Dsph [range, −6.75 to −2.25 Dsph]; Wilcoxon signed rank test, p = 0.075). In contrast, the cylindrical diopter was significantly increased in the affected eye compared to the fellow eye (−1.33 ± 1.18 diopter cylinder [Dcyl] [range, −4.75 to −0.25 Dcyl] vs. −0.83 ± 0.58 Dcyl [range, −2.00 to 0.00 Dcyl]; Wilcoxon signed rank test, p = 0.022).
Factors associated with final visual impairment
Factors associated with final visual outcomes are summarized in Tables 3 and 4. Univariable linear regression demonstrated significant associations between the following factors and final logMAR BCVA: initial logMAR BCVA (coefficient β, 0.164; 95% confidence interval [CI], 0.029 to 0.298; p = 0.019), failed autorefraction test at presentation (β = 0.588; 95% CI, 0.214 to 0.903; p = 0.003), and initial corneal epithelial defect size (β = 0.039; 95% CI, 0.011 to 0.067; p = 0.009). In model 1 of the multivariable linear regression analysis, which included the significant variables from the univariable analysis, no statistically significant differences were observed (p = 0.603, p = 0.376, and p = 0.171, respectively). However, in model 2, which included non-summer seasons, duration from symptom onset to presentation, and initial corneal epithelial defect size, the multivariable linear regression analysis showed significant associations between these variables and final logMAR BCVA (β = 0.334, p = 0.043; β = 0.092, p = 0.040; and β = 0.045, p = 0.002, respectively). In the multivariable logistic regression analysis, failed autorefraction at presentation due to an Ortho-K lens-related infectious keratitis lesion was a significant predictor of poor final visual outcome (Snellen BCVA ≤6/12; odds ratio, 38.995; p = 0.030).
Factors associated with final logMAR BCVA in the affected eye
Adjusted OR for poor final visual outcome
Discussion
We retrospectively investigated the clinical features and visual outcomes of infectious keratitis associated with Ortho-K lens in Korean pediatric population. To the best of our knowledge, this study is the largest case series report of infectious keratitis following Ortho-K lens use in Korean pediatric patients. Ortho-K lenses were initially used to correct or reduce myopia, and their use has recently increased to control myopic progression, especially in East Asia, where the prevalence of myopia is high (49.7% to 62.0%, among 12-year-old children) [19–22]. Since 2000, as the use of Ortho-K lenses has increased, and reports of infectious keratitis associated with Ortho-K lens also have risen [10,13,23–25]. As infections in children can be vision-threatening, it is important to identify risk factors for infection and factors that affect poor visual outcomes.
Previous studies have shown that in pediatric infectious keratitis patients, the use of Ortho-K lenses is up to 47% higher in low-latitude countries with higher average annual temperatures, compared to 10%–19% in relatively high-latitude countries [13,23,24]. Additionally, many studies have reported that infectious keratitis is most prevalent during the summer season [26–28]. Gorski et al. [27] reported that among 155 patients with infectious keratitis, 39% were contact lens wearers, and they observed a higher frequency of cases during the summer months (44.5%) compared to fall (12.3%), winter (21.9%), and spring (21.3%, p < 0.0001). McAllum et al. [26] found that among 41 patients with Acanthamoeba keratitis, 92.9% were contact lens wearers, and the onset of disease symptoms was significantly more prevalent in summer compared to winter (p = 0.02). In our study, the seasonal distribution of Ortho-K lens-related infectious keratitis showed that 42.3% of cases presented in summer (Fig. 2). Although Ortho-K lenses are not worn during the day, there have been reports of Acanthamoeba keratitis in Ortho-K users after swimming in contaminated water [29]. High average summer temperatures leading to the contamination of water sources or storage solutions, higher humidity as well as daytime water activities in such contaminated water, might be contributing factors.
Interestingly, in our study, the mean final logMAR BCVA was worst in fall (0.362 ± 0.443), followed by spring (0.328 ± 0.681), summer (0.225 ± 0.303), and winter (0.127 ± 0.158), although the differences were not statistically significant (Kruskall-Wallis test, p = 0.796). Despite a higher proportion of cases occurring in summer, the poor final visual outcome was significantly associated with non-summer seasons (Table 3). The frequency of infectious keratitis might have increased in the summer due to high temperatures and favorable conditions for bacterial growth; however, the non-summer seasons was associated with poorer final BCVA, possibly due to other contributing factors. There was no significant difference in initial BCVA between the summer and non-summer seasons (Mann-Whitney U-test, p = 0.561), but the initial corneal epithelial defect size was signif icantly larger in the summer (Mann-Whitney U-test, p = 0.260). This suggests that larger initial corneal epithelial defects in the summer, resulting in more severe symptoms, might lead to earlier detection and potentially better outcomes. However, the larger sample size in summer may have skewed the results, leading to better average visual outcomes in summer. Further research is needed to establish a causal relationship. There was no significant difference in the occurrence of specific pathogens by season (Fisher exact test, p = 0.230).
Watt and Swarbrick [30] reviewed 123 cases of Ortho-K related infectious keratitis and found that P. aeruginosa accounted for 37% and Acanthamoeba for 33% of the cases. In our study, the most common culture-proven causative organism was P. aeruginosa (5 of 21 cases, 23.8%), similar to findings in previous studies. Five of 26 patients (19.2%) with perineuritis were suspected of having Acanthamoeba keratitis, all of whom were culture-negative for Acanthamoeba. Among them, coinfection with S. marcescens (two patients) and Pandoraea species (one patient) was suspected. Given the very low culture-positive rate for Acanthamoeba (between 0% and 53%), a culture-negative result cannot completely rule out Acanthamoeba keratitis [31]. There are many reports of poor visual prognosis associated with Acanthamoeba keratitis following Ortho-K lens use [16,29,32–34]. In our study, all these five patients responded to medical treatment without requiring surgical intervention and there was no statistically significant difference in final logMAR BCVA between the Acanthamoeba-suspected group and the other group (0.099 ± 0.153 vs. 0.290 ± 0.445; Mann-Whitney U-test, p = 0.361). This may be attributed to the fact that four out of the five patients suspected of Acanthamoeba keratitis were initially treated with PHMB in combination with moxifloxacin from their first visit. The remaining patient, who showed a poor response to moxifloxacin before the Acanthamoeba culture results were available, had PHMB added promptly, which likely contributed to the favorable outcome. Therefore, we recommend the early initiation of PHMB treatment for lesions suspected to be Acanthamoeba keratitis, such as those presenting with perineuritis, even prior to the availability of culture results, to achieve the best outcomes.
One of 21 patients (4.8%) in our study exhibited a polymicrobial infection, with P. aeruginosa identified from corneal scraping and A. haemolyticus from the lens case. Kam et al. [35] conducted a systematic review of 172 eyes from 166 subjects diagnosed with Ortho-K–related infectious keratitis and found that 10 eyes (5.8%) had polymicrobial infections, with nine involving coinfection of P. aeruginosa with another organism. As in our study, different strains can often be reported on the cornea and lens case, which cannot completely rule out contamination; however, it can be indicative of hygiene status of lens case management and suggests that not only the cornea but also the lens case should be examined in infectious keratitis associated with Ortho-K lens. Additionally, a significant number of our patients had been using topical antibiotics before their first visit, which may explain why microorganisms were not identified in all corneal scraping cultures. In these cases, empirical treatment was administered (Table 1).
Our findings also revealed a higher incidence of S. marcescens compared to previous reports. In the systematic review by Kam et al. [35], S. marcescens was identified in only three of 140 cases (2.1%) of culture-proven Ortho-K lens-related infectious keratitis, whereas our results show a significantly higher incidence. Chen et al. [36] reported a detection rate of 5.13% for S. marcescens in contact lens case and 12.82% in hand sample from 39 Ortho-K users without infectious keratitis (mean age, 16.62 ± 7.94 years). They emphasized the importance of hand hygiene in preventing S. marcescens infections. In our study, the patients were younger (mean age, 11.9 ± 3.0 years) and potentially had poorer hand hygiene compared to those in Chen et al. [36]. To reduce S. marcescens infections, it is essential to adhere not only to proper contact lens hygiene practices but also to proper hand hygiene.
Ortho-K lenses alter the corneal shape to change refractive power, but this effect is transient, necessitating daily overnight wear. This increases the risk of infectious keratitis by reducing oxygen permeability through the contact lens, limiting eye movement that normally disrupts the microbial glycocalyx, and decreasing blinking, which helps distribute lysozyme across the corneal surface [37–39]. The primary users of Ortho-K lenses are school-aged children, who may struggle to maintain good daily contact lens hygiene and proper overnight wear, further increasing the risk of infectious keratitis. Previous studies have reported that the peak age for pediatric Ortho-K lens-related infectious keratitis is 11 to 12 years [9,10]. Our study also found that the peak age was 11 years old, with 73.1% of cases occurring in those aged 13 or younger, consistent with previous studies (Fig. 1). It could be due to difficulties in lens care among younger patients, but it may also be attributed to the fact that around age 11 years is a common period for starting Ortho-K lens use. The mean duration of Ortho-K lens wear before the onset of infection was 33.7 ± 21.2 months (range, 9 to 72 months). Among our patients, four of 26 patients (15.4%) used tap water for lens cleaning, which suggests that neglect in lens care after 2 to 3 years of Ortho-K use may have led to infectious keratitis. Therefore, careful education and management of both pediatric patients and their guardians regarding lens care are crucial.
Our study showed a female dominance, with a male to female ratio of 1:2.7, which was statistically significant (chi-square test, p = 0.029). In a previous study, Kam et al. [35] reported a female preponderance (male to female ratio, 1:1.7) from 173 eyes with Ortho-K–related infectious keratitis. Considering the pediatric population, this sex difference might be influenced by the differing behavior patterns of primary guardians responsible for lens care and wear, depending on their sex. However, further research is needed on this topic.
Chan et al. [14] reported residual central or paracentral corneal scarring in all 23 patients with Ortho-K lens-related infectious keratitis. Our study showed similar results, with 85% of cases leaving residual central or paracentral corneal scarring. Notably, the initial direction of the main corneal lesion in our study was most frequently in the temporal quadrant (56%), followed by the nasal quadrant (28%) (Fig. 3). Most residual corneal scarring occurred at the initial lesion site (Fig. 5A–5L). Since Ortho-K lenses require daily overnight wear, we speculate that horizontal rubbing during sleep or decentration towards the temporal or nasal side due to sleeping positions might trigger corneal erosion and subsequent microbiological infection. In the normal human eye, the eyeball deviates upward for 55% to 85% of sleep time during stages 2 to 4, while in stage 1 (rapid eye movement sleep), eye movements are 5% to 15% horizontal, 25% to 35% vertical, and 55% to 65% oblique, with little individual variation [40]. While it is unclear if vertical or oblique rapid eye movements cause more friction in the temporal or nasal quadrants during decentration of the Ortho-K lens, further research is needed to understand why corneal lesions frequently occur in the temporal or nasal quadrants despite the limited horizontal rapid eye movements during sleep.
Representative anterior segment photographs of severe cases. (A,B) Patient no. 2. (C,D) Patient no. 5. (E,F) Patient no. 7. (G,H) Patient no. 8. (I,J) Patient no. 15. (K,L) Patient no. 26. (A,C,E,G,I,K) At presentation. (B,D,F,H,J,L) At final follow-up.
In the analysis of 13 patients with central or paracentral opacity following infection treatment, who had refractive errors measured at their last visit, there was a tendency for hyperopic changes and increased astigmatism in the affected eye compared to the fellow eye, with the increase in astigmatism reaching statistical significance (Wilcoxon signed rank test, p = 0.022). Corneal opacity associated with Ortho-K lens-related infectious keratitis lesions may be the cause of hyperopic changes and increased astigmatism in the affected eye. Further large-scale research is needed to accurately assess changes in corneal curvature or astigmatism before and after treatment in the affected eye. Although longer follow-up is needed for our patients, we did not observe any spontaneous resolution of corneal opacities that remained after infection treatment. Notably, considering the future of these children, severe cases at presentation can result in permanent visual impairment due to corneal opacity even after complete treatment, highlighting the critical importance of preventing Ortho-K lens-related infectious keratitis (Fig. 5).
Univariable linear regression identified initial logMAR BCVA (p = 0.019), a failed autorefraction test (p = 0.003), and the initial size of the corneal epithelial defect (p = 0.009) as being significantly associated with the final logMAR BCVA. However, in model 1 multivariable linear regression analysis using these three significant independent variables, no statistically significant differences were observed (p = 0.603, p = 0.376, and p = 0.171, respectively). In model 2, which included non-summer seasons, duration from symptom onset to presentation, and initial corneal epithelial defect size, the multivariable linear regression analysis revealed significant associations between these variables and final logMAR BCVA (p = 0.043, p = 0.040, and p = 0.002, respectively). These results may be attributed to the small sample size, which makes it challenging to conduct a robust multivariable regression analysis. Variables that were not statistically significant in the univariable analysis may have shown significant results in the multivariable analysis when considered alongside other variables. Further large-scale studies are needed to identify more reliable and generalizable factors related to final visual outcomes. In the multivariable logistic regression analysis, the only significant factor for predicting poor final visual outcome (Snellen BCVA ≤6/12) was a failed autorefraction test at presentation due to an Ortho-K lens-related infectious keratitis lesion (odds ratio, 38.995; p = 0.030). The failed autorefraction test likely reflects the depth and extent of the initial central corneal lesion, which indicates the severity of the disease. Conducting an autorefraction test at the first visit could help predict the final visual outcome.
Our study has several limitations. First, it was conducted at a single tertiary center with a retrospective design, and the study population was ethnically homogeneous, which complicates the generalization of findings. Second, an appropriate control group was not included. Third, due to the retrospective nature of the study, it was not possible to determine the degree of refractive correction in the patients, limiting our ability to evaluate the relationship between degree of myopic correction and infectious keratitis. Fourth, we could not obtain the information of the fitting status of patients. Poor lens fitting can lead to corneal erosion, a risk factor for infection; however, we were unable to assess this factor. Despite this limitation, the strengths of our study include the thorough follow-up of a relatively large number of cases of Ortho-K lens-related infectious keratitis in Korean pediatric patients, which enabled us to identify factors related to final visual outcomes. In the future, we look forward to large-scale studies using confocal microscopy that could also confirm culture-negative Acanthamoeba keratitis.
In conclusion, while Ortho-K lenses offer effective myopic correction, they also carry the risk of infectious keratitis, which can result in permanent corneal opacities and potentially severe visual impairment in children. Our retrospective study revealed that 28% of children experienced poor visual outcomes, defined as a Snellen BCVA ≤6/12. Poorer outcomes were associated with factors such as delayed presentation, large corneal lesions, failure of autorefraction, and infections occurring in non-summer seasons. Given these factors, along with the higher incidence of cases among younger age groups and during the summer, it is crucial to educate children and their guardians on the proper use and maintenance of lenses. Furthermore, early detection, appropriate treatment for common pathogens, and meticulous management are essential for improving outcomes.
Acknowledgements
None.
Notes
Conflicts of Interest
None.
Funding
This work was supported by the Seoul National University Bundang Hospital Research Fund (No. 13-2020-007).