Clinical Comparison of Fusarium Keratitis according to the Initial Potassium Hydroxide (KOH) Smear: A Retrospective Study in South Korea

Purpose This study aimed to compare predisposing factors, clinical characteristics, treatment, and prognosis of Fusarium keratitis according to the result of the initial potassium hydroxide (KOH) smear. Methods This is a retrospective study of cases with Fusarium keratitis between January 2000 and December 2019 at two tertiary hospitals in South Korea. Patients were divided into two groups depending on the KOH smear result (KOH-positive and KOH-negative group), and its clinical factors were analyzed. Results Among 319 fungal keratitis, seventy-nine cases were identified with Fusarium keratitis. Forty-seven cases (59.5%) were negative in the initial KOH smear prior to their diagnosis. The most common predisposing factor for Fusarium keratitis was ocular trauma (55.7%). There were no significant differences in sex, occupation, ulcer size or shape, hypopyon, and initial visual acuity between the two groups. Differences were observed between the KOH-positive group and the KOH-negative group in terms of deep corneal infiltration (50.0% vs. 78.7%, p=0.008) and evisceration treatment (3.1% vs. 25.5%, p=0.008). The delayed time to initiate antifungal eye drops was longer in the KOH-negative group (1.13 ± 0.49 vs. 3.93 ± 4.89, p=0.002). Only the KOH-negative group combined bacterial infection. The significant risk factors for poor clinical outcomes were the central corneal lesion (odds ratio (OR) 3.50, p=0.047), a large ulcer size (size ≥ 7.5 mm2) (OR 4.98, p=0.009), and endothelial plaque (OR 7.00, p=0.031). Conclusion Initial KOH-negative patients often needed evisceration and had worse final visual outcomes. The delay of prompt initiation of antifungal treatment and combined bacterial infection result in a poor prognosis. This study highlights the initial KOH effect on early diagnosis and early treatment of Fusarium keratitis.


Introduction
Fusarium species (spp.) are ubiquitous flamentous fungi that are present in water, soil, and plant roots and are one of the most commonly reported organisms causing keratomycosis [1][2][3]. Te clinical diagnosis of Fusarium keratitis is challenging, and management is often difcult due to its insidious onset, diversity of clinical presentation, deep corneal stromal penetrating capabilities resulting in protracted response to medical therapy, prolonged microbiological identifcation time, and the emergence of broad resistance to most antifungals currently available [3][4][5]. Additionally, antifungal treatment may be delayed due to primary therapy with an antibacterial agent or antiviral agent [3]. Fusarium keratitis might progress to a deep extensive infection with perforation and endophthalmitis, which ultimately may lead to monocular blindness [1,6,7]. Several predisposing factors for the development of Fusarium keratitis include ocular trauma, particularly from vegetative or soil-contaminated objects, prior corneal surgery, previous chronic ocular surface disease, contact lens wear, and immunosuppressive disease or medication [1,4,5,8].
Te epidemiological pattern of fungal keratitis varies across geographic regions and climatic conditions [9]. Fusarium and Aspergillus species are more common in tropical and subtropical regions such as India and Mexico [9,10]. In these countries, corneal epithelial defects caused by ocular trauma are the main predisposing factor for Fusarium keratitis [5,9,10]. However, in patients using contact lenses, Fusarium keratitis has also become a problem in urban areas with temperate climates. From 2005 to 2006, contact lens-associatedFusarium keratitis outbreak was reported in Hong Kong, Singapore, France, and the United States [11][12][13][14][15][16].
Te traditional diagnostic method of fungal keratitis is mostly to stain and culture corneal specimens after scraping the ulcerative tissue. Te more accurate diagnosis of keratitis mainly relies on a fungal culture. However, the fungal culture requires longer time to identify pathogens, and generally, the cultivation time is 1-2 weeks [17]. Te potassium hydroxide (KOH) smear is one of the most commonly used methods for early direct microscopic detection of fungal flaments. One study reported that 10% KOH preparation was of marked importance, with a sensitivity and specifcity value of 99.3% and 99.1%, respectively, for detection of fungal flaments [18]. Te KOH smear is recommended in every clinic for initial rapid diagnosis of infective keratitis including fungal species. Infective keratitis can be reduced by early identifcation of microbial pathogens, which provides a prerequisite for early treatment of keratitis [17].
Terefore, we conducted a comparative study of cases with microbiologically proven Fusarium keratitis according to the result of the initial KOH smear at two tertiary hospitals in Jeolla Province, South Korea. Demography, disease history, clinical manifestations, treatment outcomes, and prognosis of patients have been analyzed retrospectively for the past 20 years (2000-2019).

Materials and Methods
We used a retrospective, linear, and comparative study design. We retrospectively reviewed medical records of all culture-proven cases of Fusarium keratitis that were treated at the Department of Ophthalmology of Jeonbuk National University Hospital and Chonnam National University Hospital in South Korea between January 2000 and December 2019. All patients underwent slit-lamp examinations and detailed clinical evaluations, including the smear and culture. Te inclusion criteria were clinical evidence of fungal keratitis, such as a positive fungal culture from a corneal specimen, positive identifcation of fungal elements on the KOH smear, and histopathology showing the presence of fungal elements. Te exclusion criteria were the cases that did not receive antifungal therapy, such as herpetic viral keratitis, Acanthamoeba infections, and noninfectious corneal diseases. Tis study was approved by the Jeonbuk National University Medical School Institutional Review Board and Ethics Committee. When studying data, patients were divided into two groups depending on the initial KOH smear: those with identifcation of fungal elements on the 10% KOH smear before the diagnosis of fungal keratitis on the culture as a KOH-positive group and those with no identifcation of fungal elements on the KOH smear before diagnosis as a KOH-negative group.
Baseline epidemiology, predisposing factors, clinical manifestations, and treatment were evaluated and compared between the KOH-positive and KOH-negative groups. Te epidemiologic characteristics included sex, laterality, occupation, residence area, and systemic diseases. Information on the past history and predisposing factors was focused on the presence of ocular trauma, use of contact lenses, preexisting chronic ocular surface disease, and prior corneal surgery. Te initial clinical manifestations included the location, shape of ulcer, depth of infltration, presence of feathery margin, satellite lesion, endothelial plaque, and hypopyon, size of corneal lesion, and presenting bestcorrected visual acuity as determined using Snellen test. Te size of the infltrate was measured by multiplying two linear dimensions and the longest linear diameter and its largest possible perpendicular within the confnes of the epithelial defect using a slit lamp [19]. Te location of ulceration was categorized into three zones, central, paracentral, and peripheral. Te depth of infltration was classifed into superfcial (anterior < 2/3 of the total corneal thickness) and deep (≥2/3 of the total corneal thickness) lesions. When hypopyon was present, its height was measured in millimeters from the inferior corneal limbus.
For laboratory diagnosis, we performed the smear and culture for each eye with suspected infectious keratitis. Corneal ulcers were scraped using a sterile scalpel blade after application of 0.5% proparacaine hydrochloride (Alcaine, Alcon, Fort Worth, TX, USA) for anesthesia. A cornealscraping specimen was taken from the margin and base of the ulcer and was placed onto glass slides for direct microscopic examination (Gram stain and 10% KOH smear) and also inoculated onto a variety of media. For the KOH smear, one drop of 10% KOH was added to the specimen and placed over the cover glass in the dermatology clinical laboratory, and the direct microscopic examination was performed by dermatologists. Tese included blood agar, chocolate agar, MacConkey agar, and Sabouraud dextrose agar plates. Samples were inoculated into Sabouraud plates for fungal detection and stored at room temperature for 28 days. Microorganisms were identifed using standard laboratory techniques that have been reported previously [20].
When a fungal infection was clinically suspected, topical and systemic antifungal treatment was started immediately. Topical antifungal therapy consisted of hourly instillation of topical 5% natamycin, or 0.2% fuconazole, or/and 0.15% amphotericin B. Systemic antifungal agents were administered with oral fuconazole (50 mg·bid·p. o.) and intravenous amphotericin B. Topical 2% voriconazole was used in case of no clinical improvement. All patients were treated topically with 3 rd or 4 th generation fuoroquinolones either alone or in a combination with fortifed topical aminoglycosides and intravenous systemic antibiotics as empirical therapy.

2
Journal of Ophthalmology Topical treatment was progressively tapered or modifed according to the clinical response and bacterial susceptibility. In cases where initial medical treatment failed, various surgical treatments such as amniotic membrane transplantation, conjunctival fap, keratoplasty, and evisceration were performed. Treatment outcomes were assessed at the end of 3 months or at the completion of treatment. Poor clinical outcomes were defned as the increase in the size of a corneal ulcer or infltrate, presence of corneal perforation, or when patients underwent penetrating keratoplasty or evisceration despite maximum medical therapy. SPSS Statistics for Windows, version 23.0 (IBM Corp., Armonk, NY, USA) was used for statistical analysis. All data are presented as a mean ± standard deviation for continuous variables and number (%) for categorical variables. Statistical analysis for normality was performed using the Kolmogorov-Smirnov test. Intergroup analysis was completed using the Mann-Whitney U test for data that were not normally distributed. Te chi-square test was used to analyze categorical variables. Pearson's correlation analysis was used to evaluate a correlation between poor clinical outcomes and variable clinical factors. A p value less than 0.05 was considered statistically signifcant.
Te biennial distribution of Fusarium keratitis is shown in Figure 1. Te peak incidence of Fusarium keratitis occurred between 2005 and 2013. However, there was no signifcant trend reported over time. Te mean time to symptom duration was 11.0 ± 15.7 days (range, 0-120 days). Accompanied systemic diseases included diabetes mellitus (15 patients, 19.0%) and hypertension (14 patients, 17.7%).
Various predisposing factors for Fusarium keratitis are identifed and summarized in Table 2. Te most common predisposing factor was ocular trauma (55.7%), followed by previous chronic ocular surface diseases (21.5%). Tere were no signifcant diferences in notable predisposing factors between the two groups. Cases with no known predisposing factors were small in number in the KOH-positive group compared to the KOH-negative group (4 vs. 15, p � 0.047). Tere were no contact lens wearers.
Among the clinical manifestations, the central location of corneal lesions was more common than the peripheral one in both groups. Also, ulcers with an irregular shape were more frequent than those with a circular shape in both groups. Tere were no signifcant diferences in the location of corneal lesions, shape of ulcers, feathery margin, satellite lesions, hypopyon, and endothelial plaques between the two groups. However, there were a higher proportion of cases with deep corneal infltration in the KOH-negative group (78.7%) than those in the KOH-positive group (50.0%) (p � 0.008).
Te mean pretreatment ulceration size was 11.1 ± 9.2 mm 2 . Tere were no signifcant diferences in the size of corneal lesions between the two groups. Initial visual acuity was 1.58 ± 0.95 logMAR in the KOH-positive group, and it was 1.53 ± 0.80 in the KOH-negative group. However, the diference was not statistically signifcant. However, impaired fnal visual acuity was higher in the KOH-negative group (1.94 ± 1.25 logMAR) than that in the KOH-positive group (1.41 ± 1.28) (p � 0.049) ( Table 3). Te delayed time to start initial antifungal eye drops was longer in the KOHnegative group (3.93 ± 4.89 days) than that in the KOHpositive group (1.13 ± 0.49 days) (p � 0.002) ( Table 3).
After treatment, the ratio of visual acuity change was a higher proportion of cases with improved visual acuity (43.8%) in the KOH-positive group compared to aggravated visual acuity (55.3%) in the KOH-negative group (Table 4).

Discussion
Few studies have reported on clinical diferences between KOH-positive and KOH-negative patients with Fusarium keratitis. In this study, 59.5% of patients were negative for the initial KOH smear prior to their fnal diagnosis of   (2), liver cirrhosis (2), tuberculosis (1), and so on.     Fusarium keratitis. Cases with deep corneal infltration and cases without signifcant predisposing factors were more frequently observed in the initial KOH-negative group. Te time delay to use initial antifungal eye drops after the frst visit was longer in the KOH-negative group. In cases of the initial KOH negative result, diagnosis of Fusarium keratitis was hindered until a defnite culture result was obtained. Also, combined bacterial infection was reported only in the KOH-negative group. In these cases, fungal keratitis could not be considered due to the initial bacterial infection. Terefore, antifungal treatment was delayed, and evisceration was more frequently performed in the KOH-negative group in this study. Geographical and seasonal factors play an important role in the distribution and epidemiology of fungal infection [9]. Fusarium spp. was found as the main causative microbes of fungal keratitis in tropical and subtropical areas such as Florida [21]. Two tertiary hospitals in this study located in Jeolla Province, south area of South Korea, subtropical area. Two hospitals were most frequently reported hospitals of fungal keratitis in South Korea, between 2008 and 2017 [22].
One study in the Netherlands between 2005 and 2016 reported a signifcant increase in the number of cases of Fusarium keratitis since 2010 [3]. However, in this study, there was no signifcant annual diference of the incidence of Fusarium keratitis between 2000 and 2019.
Te main risk factor for Fusarium keratitis was a history of ocular trauma, mainly in people that have agriculturerelated occupations [2,9,10,23]. Our study confrms this information, fnding 65.8% with agricultural occupation and 55.7% with a history of trauma, mostly soil and plant material.
With increasing contact lens wearing, Fusarium keratitis has also become a problem in urban areas with temperate climates. However, there were no patients associated with contact lens wearing in this study. One of the reasons may be that most of keratitis cases were reported in elderly patients with agricultural activity in rural areas, so they did not wear contact lenses.
Te authors predicted that patients in the KOH-positive group would have more severe clinical manifestations than those in the KOH-negative group. However, this study found no signifcant diferences in clinical characteristics between the two groups except only in terms of corneal infltration depth. At initial examination, deep corneal stromal infltration was common in the KOH-negative group. One potential explanation of this result is that mixed bacterial infection in the KOH-negative group might be more involved in deep corneal stromal infltration. In a study by Ahn et al. [24], it was reported that mixed bacterial and fungal keratitis showed more common deep corneal stromal infltration and poor prognosis. Terefore, they recommended that when treatment against the initial causative microbes of corneal ulcers fails, combined infection should be suspected and the repeated smear and culture should be considered.
Direct microscopic examination of a corneal scraping material is a rapid and available method for diagnosis of fungal keratitis [18]. Preparation of corneal scrapings with 10% KOH wet mount and trying to fnd flamentous fungi or yeast cells is a simple and inexpensive technique for diagnosis of fungal keratitis [25]. Bharathi et al. [26] investigated the diagnostic value of the 10% KOH smear for detection of fungal flaments on 3298 eyes. According to this study, the 10% KOH smear had noticeable importance with a sensitivity and specifcity positive predictive value and negative predictive value of 99.3%, 99.1%, 98.5%, and 99.6%, respectively. In comparison to these fndings, Badiee et al. [27] and Vengayil et al. [28] reported that the sensitivity and specifcity of the KOH smear was 68% and 60%, respectively. Usually, the sensitivity of the KOH smear may be adversely afected by the insufcient amount of sampling materials, the small size of the corneal ulcer, and the lack of individual experience of the microscopic observer [26]. In this study, nearly 60% patients were initially KOH negative and fnally culture positive. In the KOH-negative group, fnal visual acuity was worse than in the KOH-positive group, and also, evisceration treatment was performed more frequently. Te delay of prompt initiation of antifungal therapy in the KOHnegative group may result in a poor prognosis. Terefore, although the KOH smear and culture are of variable sensitivity, they should be performed for all corneal ulcer patients as the basic diagnostic examination. For the treatment of Fusarium keratitis, topical use of fuconazole and amphotericin B was signifcantly higher in the KOH-positive group in this study. Also, subconjunctival injection of an antifungal agent was higher in the KOHpositive group. Naturally, early diagnosis contributes to early treatment with topical and local injection of antifungal agents.
At presentation, there was no signifcant diference between the KOH-positive and KOH-negative groups in visual acuity. After treatment, improvement of visual acuity was more common in the KOH-positive group. On the contrary, cases with aggravation of visual acuity were more common in KOH-negativeFusarium keratitis. Tere was no statistically signifcant diference between the KOH-positive and KOH-negative groups in the change of visual acuity after treatment. However, at last follow-up, the KOHnegative group had signifcantly lower visual acuity. Pérez-Balbuena et al. [10] noted that fnal visual acuity after medical/surgical treatment was 64.9% patients with less hand movement (HM). Tey concluded that the outcome of Fusarium keratitis is generally poor in spite of various treatment options available.
We performed a logistic regression analysis to determine risk factors for poor clinical outcomes, in cases of such as progression, perforation, penetrating keratoplasty, and evisceration. As a result, central corneal lesions (OR 3.50, 95% CI 1.02-12.04, p � 0.047), a large ulcer size (size ≥ 7.5 mm 2 ) (OR 4.98, 95% CI 1.50-16.49, p � 0.009), and endothelial plaque (OR 7.00, 95% CI 1.19-41.27, p � 0.031) were highly associated with poor clinical outcomes. Lalitha et al. [29] reported that the presence of hypopyon was a signifcant predictor of treatment failure in fungal keratitis. Cho and Lee [30] stated signifcant risk factors for treatment failure were hypopyon and deep infltration in microbiologically proven fungal keratitis. In this study about Fusarium keratitis, deep corneal infltration and hypopyon were signifcant in univariate logistic regression analysis, but their efects were attenuated in multivariate regression analysis. Central lesion, large epithelial defect size, and endothelial plaque can be regarded as a marker of a poor prognosis.
Tis study has several limitations. First, the sample size was small. Second, this study was confned to Jeolla Province, South Korea, which is temperate to subtropical climates and with four seasons. Tird, because the study was conducted by the tertiary hospital, the results of this study cannot be generalized. Fourth, only patients with microbiological proven Fusarium keratitis were enrolled in this study, while cases of other fungal or clinically suspected fungal keratitis were excluded. Terefore, the efect of the initial KOH smear in this study cannot be generalized on all fungal keratitis. Despite such limitations, this study has an important clinical signifcance. Tis investigation highlights the initial KOH efect on early diagnosis and early treatment of Fusarium keratitis. Tis study is a clinical analysis of Fusarium keratitis in areas with high fungal prevalence in South Korea. Clinicians may use these fndings as a useful reference for various regional diferences in Fusarium keratitis.
To our knowledge, this is the frst study evaluating the clinical outcomes of Fusarium keratitis that divides two groups depending on the initial KOH smear result. In this study, the KOH-negative group had worse fnal visual acuity, and evisceration was performed frequently in the KOHnegative group than in the KOH-positive group. Te delay of prompt initiation of antifungal treatment results in a poor prognosis. Tis study highlights the initial KOH efect on early diagnosis and early treatment of Fusarium keratitis.

Data Availability
Te detailed data used to support the fndings of this study are available from the corresponding author upon request.

Conflicts of Interest
Te authors declare that they have no conficts of interest.