As conventional corneal scraping with bacterial culturing has several disadvantages such as long duration until diagnosis, invasiveness and lacks alternatives, there is demand for a novel non-invasive, rapid test in keratitis management. First aim of this study will be to evaluate the influence of TNF-alpha on Gram+ and Gram- bacteria causing keratitis. Second aim of this study will be to develop an algorithm, that should help to distinguish between different pathogens causing keratitis.

Infectious or inflammatory processes of the cornea (keratitis) are common and vision threatening. They are the 5th leading cause of blindness and especially infectious keratitis is responsible for 2,000,000 new blind eyes per year (worldwide). Commonly, the presence and nature of bacterial pathogens is determined by taking scratches from the cornea, subsequent Gram stain and examination of microbiological cultures. Major problems in therapeutic decision-making pose the long duration of bacterial culturing for pathogen detection (approximately one week), relatively low sensitivity and increasingly frequent bacterial resistances to antibiotic agents.

In small previous studies optical coherence tomography (OCT) combined with analysis of tear cytokines were found to be supportive in distinguishing between different pathogens causing keratitis. These studies appear to be a game changer in the field of rapid testing for keratitis. However, the problem of these studies is a small sample size and the use of partially outdated equipment.

Benefit/risk assessment Although there is no direct benefit for participating subjects, the results of the study can help to gain more insight into the course of keratitis and could further improve the diagnostic utilities in disease management.

There is no additional medication used in this study. The findings of the present study could help to establish a better understanding of keratitis and could enable faster and target-oriented therapy. All applied measurement techniques are painfree.

Thus, the risk to benefit ratio seems to be acceptable.

Study objectives

Primary:

• Comparison of TNF-alpha level differences in tear film (pg/ml) between Gram+ and Gram- bacteria causing keratitis (z test for non-paired samples).

Secondary:

• To develop an algorithm, that should help to distinguish between different pathogens causing keratitis using cytokine levels and OCT findings. This part of the study will be experimental (partial least squares regression and regression random forest plots)
• Levels of IL-1alpha, IL-1beta, IL-6, IL-8, IL-10, IL-17A, TNF-alpha in keratitis
• Morphological findings in optical coherence tomography: corneal thickness (µm), infiltrate thickness (µm), reflectivity of the infiltrate (grey level, where 0 is black and 255 is white).

Design: Prospective single center trial The participants will be recruited from the ambulances of Kepler University clinic, ophthalmology, Linz, Austria A total of 392 patients will be included in this study. Aimed power: 0.8 significance level after a Bonferroni correction for 7 explanatory variables (7 different cytokines): 0.07 Used values (TNF alpha 2): Group 1: 18.8 (26.36) and Group 2: 50.12 (103.83) Parameters Results Power 0.800 alpha 0.007 Mean (Group 1) 18.8 Mean (Group 2) 50.12 Std. deviation (Group 1) 26.36 Std. deviation (Group 2) 103.83 Sample size 1 146 Sample size 2 146 Power (obtained) 0.798 For correction of a potential rounding error (calculated power 0.798) one case should be added to each group.

Although there will not be any study follow-ups the investigator still expects a drop-out rate of 30% for the following reason: As published previously, 67% of all corneal biopsies result in a positive culture. 13 In those cases, where the culture does not provide positive results, the patient will have to be excluded from the analysis.

Therefore, the investigator will include 147 + 147 + (147+147)*0.33 = 391.02 eyes -> 392 eyes Analysis of data is only possible, if the corneal biopsy is positive (a pathogen causing the keratitis is found). Therefore, the investigator expects that approximately 30% of the patients will not be included for analysis. In these cases, the OCT images will still be evaluated, but the tear film sample will be either disposed according to the protocol of tissue waste or if the investigating ophthalmologist expects a non-infectious type of keratitis (such as a neurotrophic ulcer) the cytokine quantification will be performed.

Measurement techniques CASIA 2 (Tomey, Japan) is a ss-OCT biometer that allows high resolution imaging of the anterior segment of the eye.

Tear sampling and cytokine analysis Samples will be collected by conjunctival lavage of the affected eye without topical anesthesia. Sterile normal saline will be infused at room temperature into the lower conjunctival sac by gently pulling down the lower eyelid and with a needle-free insulin syringe. After few seconds, the lavage fluid will be aspirated back into the same syringe without the syringe touching the conjunctiva. Afterwards, the sample will be stored at -80° Celsius until all samples are collected. As mentioned in those samples, where the corneal biopsy was negative, will either be disposed, or (if the ophthalmology expects a non-infectious type of keratitis) still be analysed. Therefore, the conjunctival fluid from the syringes will be pooled and spun down in a centrifuge (400xg for 10 min at 4°C). The Supernatants will be analyzed using the pro-inflammatory multiplex assay kit from Meso Scale Discovery (Gaithersburg, MD, USA) for levels of the following cytokines and chemokines: IL-1alpha, IL-1beta, IL-6, IL-8, IL-10, IL-17A, TNF-alpha. Eighty samples will be analysed at the same time.

Inclusion Criteria:

• Men and women age of at least 18 years to 119 years
• Written informed consent prior to study-related procedures
• First presentation with keratitis (corneal epithelial ulceration, corneal edema and/or stromal infiltration with inflammatory cells) within 10 days.
• Untreated keratitis or keratitis not responding to the given treatment

Exclusion Criteria:

Subjects will be excluded if one or more of the following criteria apply:

• Symptoms for more than 10 days
• Presence or history of a severe medical condition as judged by the clinical investigator
• Pregnancy, planned pregnancy or lactating

The participants will be recruited from the ambulances of Kepler University clinic, ophthalmology, Linz, Austria

If IPD is going to be shared, is still in evaluation and will be decided throughout the study.