Abstract

   Since the environmental, behavioral, and socioeconomic factors in the
   etiology of keratoconus (KTCN) remain poorly understood, we
   characterized them as features influencing KTCN phenotype, and
   especially affecting the corneal epithelium (CE). In this case-control
   study, 118 KTCN patients and 73 controls were clinically examined and
   the Questionnaire covering the aforementioned aspects was completed and
   then statistically elaborated. Selected KTCN-specific findings were
   correlated with the outcomes of the RNA-seq assessment of the CE
   samples. Male sex, eye rubbing, time of using a computer after work,
   and dust in the working environment, were the substantial KTCN risk
   factors identified in multivariate analysis, with ORs of 8.66, 7.36,
   2.35, and 5.25, respectively. Analyses for genes whose expression in
   the CE was correlated with the eye rubbing manner showed the enrichment
   in apoptosis (TP53, BCL2L1), chaperon-related (TLN1, CTDSP2, SRPRA),
   unfolded protein response (NFYA, TLN1, CTDSP2, SRPRA), cell adhesion
   (TGFBI, PTPN1, PDPK1), and cellular stress (TFDP1, SRPRA, CAPZB)
   pathways. Genes whose expression was extrapolated to the allergy status
   didn’t contribute to IgE-related or other inflammatory pathways.
   Presented findings support the hypothesis of chronic mechanical corneal
   trauma in KTCN. Eye-rubbing causes CE damage and triggers cellular
   stress which through its influence on cell apoptosis, migration, and
   adhesion affects the KTCN phenotype.

Introduction

   Keratoconus (KTCN) is a bilateral corneal disease characterized by
   progressive thinning and scarring of the cornea, iron deposition, and
   breaks in Bowman’s layer [[34]1]. The usage of state-of-the-art
   technologies extends this definition to include detected abnormalities
   in both posterior corneal elevation and corneal thickness distribution,
   as well as in the corneal epithelium thickness profile [[35]2, [36]3].
   Compound myopic astigmatism is the most prevalent refractive error in
   the manifest refraction in KTCN patients, followed by hyperopic
   compound, mixed and myopic astigmatism [[37]4]. Visual function support
   is provided with eyeglasses, contact lenses, or intrastromal corneal
   ring segments (ICRS) [[38]5]. Progression of KTCN can be
   modified–slowed down with corneal cross-linking procedure (CXL), which
   normalizes corneal parameters and improves visual acuity [[39]6].
   Penetrating keratoplasty, deep anterior lamellar keratoplasty, and
   Bowman layer transplantation are treatment options in the advanced
   stages of the disease [[40]5].

   KTCN affects all ethnic groups worldwide and the disease prevalence
   varies between the described populations. One of the lowest incidence
   rates (0.0068%) was reported in North Macedonia [[41]7] and the highest
   (4.79%) in pediatric patients in Saudi Arabia [[42]8]. Based on the
   performed meta-analysis, the frequency of 1.3/1000 in the general
   population was estimated [[43]9].

   The majority of patients are diagnosed with KTCN at an early age,
   usually in the second decade of life [[44]10]. In the youngest
   patients, under 10 years of age, KTCN has been reported with frequent
   eye rubbing and eye allergies [[45]11, [46]12].

   According to the Global consensus on keratoconus and ectatic diseases
   [[47]2], Down syndrome, connective tissue disorders (Marfan syndrome,
   Ehlers-Danlos syndrome), Leber congenital amaurosis, floppy eyelid
   syndrome, Asian and Arabian ethnicity, KTCN diagnosed in family
   members, ocular allergy, atopy, and the eye rubbing as an exclusive
   behavior factor, are KTCN risk factors [[48]2]. In addition to the
   listed genetic syndromes, familial aggregation [[49]13], discordance
   between twins [[50]14], identified numerous KTCN loci, candidate genes
   and sequence variants [[51]13], and higher KTCN frequency in offspring
   in communities with high consanguinity rates, confirm the genetic
   background in KTCN [[52]9, [53]15]. Therefore, a genetic predisposition
   together with environmental factors could trigger a biochemical cascade
   influencing the disease onset [[54]16].

   As aspects of the influence of environmental factors on KTCN onset
   remain a controversial and contentious subject in KTCN research, the
   aim of this study was to characterize various environmental/behavioral
   and socioeconomic factors potentially influencing the KTCN phenotype,
   and especially the corneal surface. We then correlated the
   clinical/behavioral identified outcomes with the relevant data from
   analyses of transcriptomic features found in the same assessed patients
   to further characterize the revealed relationships.

Materials and methods

Ophthalmic examination and patients’ inclusion and exclusion criteria

   This single-center (Optegra Eye Health Care Clinic in Poznan, Poland),
   prospective, case-control study was conducted from October 2019 to May
   2022. The study protocol was approved by the Institutional Review Board
   at Poznan University of Medical Sciences, Poznan, Poland (resolutions
   no. 200/22). Subject recruitment, sample collection, and other
   performed procedures/methods were conducted following the institutional
   guidelines and regulations, and in accordance with the Declaration of
   Helsinki. The written informed consent, for study participation and
   publication without identifying information, before subject recruitment
   from subjects and/or their legal guardian(s) was obtained. The
   individual whose photographs are presented in [55]Fig 1 in this
   manuscript has given written informed consent (as outlined in PLOS
   consent form) to publish the case details.

Fig 1. The identification of the eye rubbing manner in the examined
individuals.

   [56]Fig 1
   [57]Open in a new tab

   The photographs, numbered from 1 to 8, ranked ascending by their
   severity/potential harmfulness to the eyeball, included in the
   Questionnaire for the indication of the eye rubbing manner in the
   examined individuals. In the Questionnaire, the aspect of eye rubbing
   manner was verified using the multiple question method (Study
   Questionnaire, question no. 9: ‘Please select the photo that most
   closely resembles the way you rub your eyes’).

   Each individual underwent a complete ophthalmological examination,
   including the assessments of both uncorrected (UDVA) best-corrected
   distance visual acuity (BDVA), intraocular pressure (IOP), corneal
   tomography with rotating Scheimpflug camera Pentacam (Oculus
   Optikgeraete GmbH, Wetzlar, Germany), epithelial thickness mapping
   MS-39 (CSO, Costruzione Strumenti Oftalmici, Florence, Italy),
   slit-lamp and dilated funduscopic examination. Additionally, towards
   screening of the dry eye syndrome symptoms the Ocular Surface Disease
   Index (OSDI) was evaluated and in suspected individuals, the Tear
   Breakup Time (TBUT) was checked.

   The inclusion criteria for patients with KTCN were the diagnosis of
   KTCN, the following indices were analyzed: Belin-Ambrósio Enhanced
   Ectasia Display Total Deviation Value, Ambrósio relational
   thicknessvalues, pachymetric progression indices, the index of vertical
   asymmetry, anterior and posterior elevation [[58]17], and corneal
   epithelial thickness profile [[59]3].

   The experienced ophthalmologists subspecialised in cornea and
   refractive surgery rated all corneal images.

   The group of controls consisted of volunteers or regular clinic
   patients or KTCN family members, who agreed on a long detailed clinical
   examination followed by a detailed Questionnaire. The inclusion
   criteria for control individuals included no clinical evidence/symptoms
   of KTCN.

   Exclusion criteria for both studied groups were co-occurring severe
   ophthalmic condition and genetic diseases/disorders. Additionally, the
   ongoing ocular surface inflammation was the exclusion criterion.

Collection of environmental/behavioral/socioeconomic data

   A detailed Questionnaire, embracing questions on demographics,
   behavioral, environmental and socioeconomic aspects, was completed by
   each participant, in the presence of the researcher. The place of
   living up to the age of 15 (village or city up to 20,000 residents or
   city from 20,000 to 100,000 residents or city from 100,000 to 500,000
   residents or city with over 500,000 residents); education level
   (primary or vocational education or high school or university); time of
   outdoor physical activity up to the age of 15; previous conjunctivitis,
   dry eye syndrome, eye trauma, contact lenses use, ocular diseases,
   allergy (drug/food/pollens/dust/grass), asthma, atopic dermatitis,
   thyroid gland diseases, genetic diseases, smoking, past pregnancies
   were recorded.

   Detailed questions regarding eye rubbing manner of study individuals
   embracing frequency of this behavior (occasionally or sometimes or
   often), dominant hand used, eye preference, preference of a part of
   hand used to rub, preference of a part of eye to rub, and details about
   the eye rubbing right after waking up were asked. Additionally, the 8
   photographs showing different manners of eye rubbing and ranked
   ascending by their severity/potential harmfulness to the eyeball
   (ordinal scale) were presented to the examined person. The order of the
   photos was determined based on the expertise of ophthalmologists who
   performed the clinical examination and the researcher processing the
   surveys.

   Also, aspects of professional occupation, presence of dust in working
   environment (defined as an exposition to filings of metals/building
   materials/construction materials or excessive dust), excessive UVA
   exposure, leisure activities (hobbies), reading habits, time spent in
   front of a screen (computer/tablet/smartphone/electronic
   reader/television set), hormones intake and sleeping habits (body
   position, head orientation and hand/arm position) were evaluated.

Statistical analyses of clinical parameters and environmental/ behavioral/
socioeconomic data

   The JASP Software 5 [[60]18] was used in statistical analyses. The
   normality of the continuous data was assessed by Kołomogorov-Smirnov
   and Shapiro-Wilk tests. The t-test was applied for continuous variables
   with a normal distribution. If the normality assumption was not
   satisfied, the Mann–Whitney test was applied. In the case of nominal
   data, the Chi-square test was executed.

   The odds ratio on univariate logistic regression with 95% confidence
   intervals were calculated. The multivariate logistic regression was
   used to create the logistic model of KTCN risk factors, where odds
   ratios and 95% confidence intervals for variables were recalculated.
   The confusion matrix (performance diagnostic) of final model was
   computed, followed by performance metrics including accuracy, area
   under curve, sensitivity, specificity, and precision. The ROC curve was
   prepared.

   For all performed statistical tests p-values ≤0.05 were considered as
   statistically significant.

Integrative analyses and biological interpretation

   The integration of the obtained clinical and
   environmental/behavioral/socioeconomic data with the corresponding
   transcriptomic experimental data derived from the RNA-seq investigation
   of the cone (central) region of CE was performed. The transcriptomic
   assessment of the samples, embracing sample preparation, sequencing,
   and bioinformatic analyses were performed as previously [[61]19].
   Briefly, the CE samples were obtained during CXL [[62]20] and
   photorefractive keratectomy (PRK) [[63]21] procedures performed in KTCN
   patients and control individuals undergoing refractive error
   correction, respectively. The cone (central) region of each CE sample
   was assessed manually (based on the corneal tomography and epithelial
   thickness mapping), positioned in the intended orientation on the
   microscope slide with the use of light microscope and previously made
   stamps, and separated. Separated CE samples were transferred from the
   microscope slides to the lysis solution (Norgen Biotek, Thorold, ON,
   Canada) and total RNA, DNA, and proteins were extracted according to
   the instructions supplied with the RNA/DNA/Protein Purification Plus
   Micro Kit (Norgen Biotek). Total RNA libraries were prepared according
   to a previously established protocol [[64]19], using TruSeq Stranded
   Total RNA Library Prep Gold (Illumina, San Diego, CA, USA) in
   accordance with the manufacturer’s protocol. A 100-bp paired-end
   sequencing run was performed on a NovaSeq 6000 platform (Illumina). The
   CE samples were sequenced with an average coverage of 100 million read
   pairs per sample. Bioinformatic analyses were executed according to a
   previously established protocol [[65]19].

   The relations between expression of particular genes (in TPM,
   transcript per million) and selected
   environmental/behavioral/socioeconomic data was executed using Pearson
   correlation or t-test. The set of dysregulated genes was analyzed using
   ShinyGO [[66]22], and p-value ≤ 0.05 and/or FDR ≤ 0.05 were the cutoffs
   for pathway enrichment analyses. The Reactome database [[67]23] was
   used as a pathway term source.

   For visualization of the integrative analyses the ggplot2 package
   [[68]24] in Rstudio (v. 4.1.3) [[69]25] was used.

Results

Clinical characteristics of patients and controls

   The 118 patients with KTCN (21 females and 97 males) and 73 control
   individuals (42 females and 31 males) were involved in this study.
   Among examined control individuals, no signs of KTCN have been
   disclosed. Clinical characteristics, including UDVA, BDVA, flat
   keratometric readings (K1), steep keratometric readings (K2), maximum
   simulated keratometry (Kmax), anterior and posterior elevation of the
   cornea, thinnest corneal thickness (TCT), thinnest epithelial
   thickness, axial length (AL), and IOP data of the examined study groups
   are compiled in [70]Table 1. All presented clinical data differed
   significantly between the studied groups. No statistical differences in
   clinical parameters such as K1, K2, Kmax, anterior and posterior
   elevations, and TCT, have been found in comparison between male and
   female patients with KTCN, and between adult and adolescent patients
   with KTCN ([71]S1 and [72]S2 Tables).

Table 1. Clinical characteristics of the examined patients with KTCN and
control individuals.

   The statistically significant differences, p ≤0.05 or p <0.001, are
   indicated.
   KTCN (n = 118) Controls (n = 73)
   x ± SD Median x ± SD Median
   UDVA OD 0.46 ± 0.37 0.30[73]^‡ 0.74 ± 0.46 0.85
   BDVA OD 0.80 ± 0.35 0.90[74]^‡ 1.12 ± 0.20 1.10
   UDVA OS 0.44 ± 0.35 0.30[75]^‡ 0.79 ± 0.46 0.95
   BDVA OS 0.78 ± 0.34 0.90[76]^‡ 1.14 ± 0.14 1.10
   K1 OD [D] 45.35 ± 5.14 43.95[77]^‡ 42.93 ± 1.34 42.90
   K2 OD [D] 48.07 ± 6.11 46.10[78]^‡ 43.86 ± 1.47 43.90
   Kmax OD [D] 53.93 ± 9.53 51.35[79]^‡ 44.28 ± 1.53 44.30
   anterior elevation OD [μm] 23.55 ± 18.42 18.00[80]^‡ 2.81 ± 1.77 2.00
   posterior elevation OD [μm] 50.78 ± 37.95 39.50[81]^‡ 6.60 ± 4.83 6.00
   TCT OD [μm] 465.34 ± 56.02 471.00[82]^‡ 542.49 ± 35.51 544.00
   thinnest epithelial thickness OD [μm] 43.16 ± 5.45 44.00[83]^‡ 48.59 ±
   3.81 48.00
   K1 OS [D] 45.45 ± 4.71 43.75[84]^‡ 43.01 ± 1.40 43.00
   K2 OS [D] 48.32 ± 5.63 46.45[85]^‡ 43.87 ± 1.59 44.10
   Kmax OS [D] 54.54 ± 9.16 52.10[86]^‡ 44.30 ± 1.58 44.30
   anterior elevation OS [μm] 26.11 ± 18.40 22.50[87]^‡ 2.82 ± 1.99 2.00
   posterior elevation OS [μm] 52.72 ± 32.55 49.50[88]^‡ 6.44 ± 4.15 5.50
   TCT OS [μm] 470.64 ± 53.55 474.50[89]^‡ 541.93 ± 35.21 542.00
   thinnest epithelial thickness OS [μm] 43.52 ± 5.67 44.00[90]^‡ 49.33 ±
   4.08 49.00
   AL OD [mm] 24.00 ± 0.89 23.95[91]^† 23.57 ± 0.98 23.40
   AL OS [mm] 23.92 ± 0.88 23.85[92]^† 23.55 ± 0.92 23.40
   IOP OD [mmHg] 13.17 ± 3.18 13.00[93]^‡ 16.46 ± 2.62 16.00
   IOP OS [mmHg] 13.13 ± 3.30 13.00[94]^‡ 16.54 ± 2.39 16.00
   KTCN in a family member n = 18 (15.25%)[95]^† n = 4 (5.48%)
   Any eye disease in a family member n = 60 (50.85%) n = 29 (39.73%)
   [96]Open in a new tab

   Abbreviations and symbols in the Table: x–average, SD–standard
   deviation, OD–oculus dexter, OS—oculus sinister, UDVA–uncorrected
   distance visual acuity, BDVA–best-corrected distance visual acuity, K1
   –flat keratometric readings, K2 –steep keratometric readings,
   Kmax–maximum simulated keratometry, TCT–thinnest corneal thickness,
   AL—Axial length, IOP–intraocular pressure

   ^† indicates statistically significant differences between KTCN
   patients and controls with p-value ≤0.05

   ^‡ indicates statistically significant differences between KTCN
   patients and controls with p-value <0.001

   The familial occurrence of KTCN concerned 11 patients with KTCN,
   including two pairs of siblings (brother-brother and sister-sister),
   two pairs of fathers and their sons, and one trio of an affected father
   with two daughters with KTCN.

Behavioral, environmental, and socioeconomic aspects in KTCN

   The outcomes of analyzes of selected behavioral, environmental, and
   socioeconomic aspects evaluated in this study are presented in Tables
   [97]2 and [98]3, together with their statistical significance. All
   analysed data are compiled in [99]S3 Table, while additional results of
   comparison of male and female patients with KTCN and adult and
   adolescent patients with KTCN in aspect of selected behavioral,
   environmental, and socioeconomic factors are presented in [100]S4 and
   [101]S5 Tables, respectively. Studied groups have differed in the
   aspects of age, sex, and place of living, with a predominance of
   younger males living in villages and minor cities in the KTCN group.

Table 2. Qualitative and quantitative data/results of analyzes of selected
behavioral, environmental, and socioeconomic aspects.

   Variables KTCN (n = 118) Control (n = 73) p-value
   Age (years), mean±SD 27.3 ± 8.7 31.5 ± 10.7 0.004
   Sex < 0.001
   Female 21 (17.78%) 42 (57.53%)
   Male 97 (82.20%) 31 (42.47%)
   Level of education 0.002
   Primary 14 (11.87%) 2 (2.74%)
   Vocational education 11 (9.32%) 1 (1.37%)
   High school 51 (43.22%) 27 (36. 99%)
   University 42 (35.59%) 43 (58.90%)
   Place of living up to the age of 15 0.004
   Village 43 (36.44%) 32 (44.44%)
   City up to 20000 residents 19 (16.10%) 17 (23.61%)
   City from 20000 to 100000 residents 7 (5.93%) 11 (11.11%)
   City from 100000 to 500000 residents 30 (25.43%) 3 (4.17%)
   City with over 500000 residents 19 (16.10%) 12 (16.67%)
   Allergy 0.128
   Yes 45 (38.14%) 20 (27.40%)
   No 73 (61.86%) 53 (72.60%)
   Food Allergy 0.223
   Yes 8 (6.78%) 2 (2.74%)
   No 110 (93.22%) 71 (97.26%)
   Pollen/grass/dust Allergy 0.114
   Yes 42 (35.59%) 18 (24.66%)
   No 76 (64.41%) 55 (75.34%)
   Professional occupation 0.128
   Student 27 (23.68%) 9 (12.50%)
   Non-office worker 55 (48.25%) 36 (50.00%)
   Office worker 32 (28.07%) 27 (37.50%)
   Dust in the working environment 0.016
   Yes 34 (28.81%) 10 (13.70%)
   No 84 (71.19%) 63 (86.30%)
   Using a computer at work (hours per day) 5.7±2.8 6.1±2.3 0.617
   Using a computer after work (hours per day) 2.4±1.2 1.6±0.9 < 0.001
   [102]Open in a new tab

   Note: the respondents did not always answer all the questions,
   therefore the summation of the answers in individual questions in the
   table may be incomplete.

Table 3. The data/results of analyzes of the eye rubbing behavior.

   Variables KTCN (n = 118) Control (n = 73) p-value
   Eye rubbing < 0.001
   Yes 109 (92.37%) 50 (68.49%)
   No 9 (7.62%) 23 (31.51%)
   Frequent eye rubbing 0.011
   Yes 10 (8.48%) 0 (0.00%)
   No 108 (91.52%) 73 (100.00%)
   Dominant hand 0.006
   Right 100 (84.75%) 71 (97.26%)
   Left 18 (15.25%) 2 (2.74%)
   More frequently rubbed eye 0.020
   Both 79 (72.48%) 44 (1.67%)
   Right 14 (12.84%) 3 (6.25%)
   Left 16 (14.68%) 1 (2.08%)
   Part of the hand used for rubbing 0.052
   Fingertips 37 (41.11%) 23 (48.94%)
   Base of hand 2 (2.22%) 3 (6.38%)
   Knuckles 29 (32.22%) 18 (38.30%)
   Fists 22 (24.45%) 3 (6.38%)
   Eye rubbing with a fist 0.009
   Yes 22 (24.44%) 3 (6.38%)
   No 68 (75.56%) 44 (93.62%)
   The upper eyelid as the most frequently rubbed part 0.033
   Yes 45 (38.14%) 17 (23.29%)
   No 73 (61.86%) 56 (76.71%)
   The lower eyelid as the most frequently rubbed part 0.004
   Yes 40 (33.90%) 11 (15.07%)
   No 78 (66.10%) 62 (84.93%)
   Type of eye rubbing indicated in response to presented photographs
   0.075
   Photography no. 1 11 (11.96%) 14 (28.00%)
   Photography no. 2 14 (15.22%) 11 (22.00%)
   Photography no. 3 9 (9.78%) 4 (8.00%)
   Photography no. 4 7 (7.61%) 7 (14.00%)
   Photography no. 5 10 (10.87%) 4 (8.00%)
   Photography no. 6 9 (9.78%) 2 (4.00%)
   Photography no. 7 6 (6.52%) 2 (4.00%)
   Photography no. 8 26 (28.26%) 6 (12.00%)
   Photographs no. 1–4 or 5–8 0.002
   Photography 1 or 2 or 3 or 4 41 (44.56%) 36 (72.00%)
   Photography 5 or 6 or 7 or 8 51 (55.44%) 14 (28.00%)
   Rubbing the eyes immediately after waking up 0.191
   Yes 50 (42.37%) 24 (32.88%)
   No 68 (57.63%) 49 (67.12%)
   [103]Open in a new tab

   Note: the respondents did not always answer all the questions,
   therefore the summation of the answers in individual questions in the
   table may be incomplete.

   Substantial differences were found in Questionnaire answers concerning
   the manner of eye rubbing ([104]Table 3). In the multiple questions,
   the KTCN group more often admitted to the eye rubbing and indicated the
   fist as part of the hand used for the eye rubbing. Moreover, the
   patients rubbed most of the eye surface, upper and/or lower eyelid in
   contrast to the inner and/or outer eye corner rubbed in controls. These
   behaviors were confirmed by the more frequent selection of the
   photographs no. 5–8 (as presented in [105]Fig 1) from the remaining
   photographs included in question no. 9 in the Questionnaire. While we
   did not found differences in the manner of eye rubbing between adult
   and adolescent patients with KTCN, there was a significant difference
   considering sex of patients, that is male patients more frequently used
   knuckles, a base of hand, all fingers and/or a fist (as presented in
   photographs no. 5–8) ([106]S4 and [107]S5 Tables). Moreover, in the
   KTCN group the patients were more frequently left-handed.

   There were no differences in the frequency of reported conjunctivitis,
   dry eye syndrome, eye trauma, asthma, atopic dermatitis, and allergy to
   neither food nor pollens/grass/dust, between the studied groups
   ([108]S3 Table).

   Regarding the working environment, despite the lack of difference in
   the type of professional occupation between studied groups, we found
   that the presence of dust in the working environment was more
   frequently indicated by patients with KTCN ([109]Table 2), and
   especially male patients with KTCN ([110]S4 Table). Moreover, by
   assessing responses to questions regarding leisure activities, we found
   that KTCN patients spend more time using a computer after work and a
   smartphone in the dark ([111]Table 2). There was no difference in the
   declared hobbies ([112]S3 Table).

   Finally, considering the sleeping position including head orientation
   and hand/arm position, no differences were found comparing the studied
   groups ([113]S3 Table).

   Additional statistics regarding presence of dust in the working
   environment and eye rubbing, most frequently rubbed eye and right /
   left-handedness, most frequently rubbed eye and more affected eye, eye
   rubbing and absolute differences in TCT between left and right eye, eye
   rubbing and allergy status, and eye rubbing and time spent using a
   computer are presented in [114]S6 Table.

Risk factors for KTCN in univariate analyses

   The results of risk analysis of behavioral, environmental, and
   socioeconomic factors in KTCN, presented as the odds ratio, 95%
   confidence interval (CI), and p-value are compiled in [115]Table 4.
   Statistical significance was found for 12 analyzed variables. The male
   sex (odds ratio 6.26), living in the median size city (100,000–500,000
   residents) (OR 6.32), eye rubbing (OR 5.57), frequent rubbing of the
   left eye (OR 8.91), left handedness (OR 6.39), eye rubbing with fist
   (OR 4.75), rubbing of both eyes simultaneously with all fingers
   (photography no. 6; OR 5.73) or fists (photography no. 8; OR 5.52),
   rubbing of the most of the eye surface with the use of greater force
   (rubbing with knuckles/base of hand /all fingers /fists, photographs
   no. 5–8; OR 3.20), presence of dust in the working environment (OR
   2.55), and longer time of using a computer after work (OR 1.91) were
   factors recognized as significantly increasing the KTCN risk. Higher
   education (university; OR 0.52) was the only variable reducing the risk
   of KTCN.

Table 4. Selected results of the risk analysis of behavioral, environmental
and socioeconomic factors in KTCN.

   For each variable, the odds ratio, 95% confidence interval (CI), and
   p-value are presented.
   Variable Odds Ratio 95% CI p-value
   Sex (male) 6.26 3.23–12.13 <0.001
   Level of education
       University (vs high school) 0.52 0.28–0.97 0.041
   Place of living up to the age of 15
       City with 100,000–500,000 residents (vs city with more than 500,000
   residents) 6.32 1.57–25.34 0.009
   Allergy (yes) 1.63 0.87–3.08 0.130
       Food allergy (yes) 2.58 0.53–12.51 0.239
       Pollen/grass/dust allergy (yes) 1.69 0.88–3.24 0.115
   Asthma (yes) 1.60 0.48–5.29 0.444
   Smoking (yes) 0.86 0.421–1.77 0.686
   Eye rubbing (yes) 5.57 2.41–12.91 <0.001
   Dominant hand (left) 6.39 1.44–28.41 0.015
   More frequently rubbed eye
       Left (vs both) 8.91 1.14–69.47 0.037
   Eye rubbing with fists (yes) 4.75 1.34–16.80 0.016
   Rubbing of the upper eyelid (yes) 1.39 0.70–2.79 0.349
   Rubbing of the lower eyelid (yes) 2.08 0.96–4.50 0.064
   Type of eye rubbing indicated in response to presented photographs
       Photography no.6 (vs No.1) 5.73 1.02–32.10 0.047
       Photography no.8 (vs No.1) 5.52 1.68–18.10 0.005
       Photographs no. 1–4 or 5–8 (vs No. 5–8) 3.20 1.52–6.72 0.002
   Working environment with dust (yes) 2.55 1.17–5.55 0.018
   Time of using a computer at work (hours per day) 0.95 0.82–1.09 0.438
   Time of using a computer after work (hours per day) 1.91 1.31–2.77
   <0.001
   Using a smartphone (yes) 2.21 0.79–6.22 0.133
   Time of using a smartphone during the day (hours per day) 1.20
   0.93–1.54 0.154
   Using a smartphone during the night (yes) 0.88 0.47–1.62 0.671
   Time of using a smartphone during the night (hours per day) 1.51
   0.85–2.68 0.157
   [116]Open in a new tab

   In addition, in [117]Table 4 the non-significant but previously
   discussed factors (such as allergy or asthma) are presented.

Multivariate analysis and the resulting KTCN risk factor model

   Male sex, eye rubbing, time of using a computer after work, and dust in
   the working environment, were the substantial KTCN risk factors
   identified in multivariate analysis, with OR values of 8.66, 7.36,
   2.35, and 5.25, respectively. The parameters of the recognized model,
   including the confusion matrix and performance metrics (accuracy, Area
   Under the Curve (AUC), sensitivity, specificity, and precision) are
   presented in [118]Table 5A–5C, whereas the Receiver Operating
   Characteristic (ROC) curve for these variables in the logistic model,
   is shown in [119]Fig 2.

Table 5. Substantial KTCN risk factors, odds ratios, the confusion matrix,
and performance metrics in a multivariate analysis.

   A. Odds ratios
   Variable Odds Ratio 95% CI p-value
   Sex (male) 8.66 3.16–23.79 <0.001
   Eye rubbing (yes) 7.36 2.02–26.89 0.003
   Time of using a computer after work (hours per day) 2.35 1.42–3.88
   <0.001
   Dust in the working environment (yes) 5.25 1.23–22.33 0.025
   B. Confusion matrix
   Predicted
   Observed Control KTCN % Correct
   Control 39 14 73.585
   KTCN 11 67 85.897
   Overall % Correct 80.916
   C. Performance metrics
   Value
   Accuracy 0.809
   AUC 0.882
   Sensitivity 0.859
   Specificity 0.736
   Precision 0.827
   [120]Open in a new tab

Fig 2. ROC curve of multivariate model of KTCN risk factors.

   [121]Fig 2
   [122]Open in a new tab

   ROC curve for the multivariate logistic regression analysis
   illustrating the performance of KTCN risk factors model. Model
   statistic values are presented in [123]Table 5.

Effect of eye rubbing on corneal surface transcriptome and biological
interpretation

   The results of the integration of transcriptomic data of the cone
   (central) region of the corneal epithelium (CE) of 23 KTCN patients and
   five controls and chosen environmental and behavioral factors are
   presented in Figs [124]3–[125]5 and [126]Table 6.

Fig 3. Results of pathway enrichment analysis.

   [127]Fig 3
   [128]Open in a new tab

   Results of pathway enrichment analysis for genes which expression in
   corneal epithelium was found to be correlated with the eye rubbing
   manner. The type of eye rubbing was indicated in response to presented
   photographs, included in [129]Fig 1. Transcriptome data was obtained
   for males and females from both study groups. The Reactome database was
   chosen as the source of the pathways.

Fig 5. Results of Pearson’s correlation.

   [130]Fig 5
   [131]Open in a new tab

   Pearson’s correlation results between the expression of selected genes,
   TRIM8, JUNB, and LTBP3 (in TPM), in the cone (central) region of
   corneal epithelium and time spent using a computer after work. Values
   of regression and statistical significance are presented in plots.

Table 6. The results of the pathway enrichment analysis.

   The results of the pathway enrichment analysis for genes whose
   expression in the cone region of corneal epithelium was found to be
   correlated with the eye rubbing manner.
   Pathway Fold Enrichment Enrichment FDR nGenes Pathway Genes Genes
   Asparagine N-linked glycosylation 3.5712 0.0126 14 305 CTSA, CAPZB,
   ARFGAP1, NAPA, ST3GAL4, GORASP1, TBC1D20, TMEM115, PRKCSH, B4GALT3,
   DCTN5, LMAN2, MAN1B1, DYNLL2
   Transport to the Golgi and subsequent modification 4.6261 0.0126 11 185
   CAPZB, ARFGAP1, NAPA, ST3GAL4, GORASP1, TBC1D20, TMEM115, B4GALT3,
   DCTN5, LMAN2, DYNLL2
   XBP1S activates chaperone genes 9.7253 0.0126 6 48 ARFGAP1, WFS1,
   KLHDC3, TLN1, CTDSP2, SRPRA
   IRE1alpha activates chaperones 8.8078 0.0126 6 53 ARFGAP1, WFS1,
   KLHDC3, TLN1, CTDSP2, SRPRA
   Membrane Trafficking 2.4349 0.0214 21 671 DVL2, AP2B1, CAPZB, PACSIN2,
   AP1B1, ARFGAP1, STX4, STX10, ANKRD27, NAPA, GORASP1, TBC1D20, TMEM115,
   KIF1C, VPS4A, DCTN5, YWHAB, LMAN2, ALS2CL, AP1G2, DYNLL2
   Intrinsic Pathway for Apoptosis 7.6527 0.0214 6 61 TP53, YWHAB, STAT3,
   BCL2L1, TFDP1, DYNLL2
   ER to Golgi Anterograde Transport 4.5469 0.0214 9 154 CAPZB, ARFGAP1,
   NAPA, GORASP1, TBC1D20, TMEM115, DCTN5, LMAN2, DYNLL2
   Post-translational protein modification 1.8475 0.0242 36 1516 SCMH1,
   CTSA, OTUD5, CAPZB, PIGS, JOSD1, ARFGAP1, AXIN1, NAPA, TNKS2, DPH1,
   WFS1, ST3GAL4, GORASP1, TBC1D20, TMEM115, PRKCSH, DDA1, UBE2M, TRIM28,
   CKAP4, TP53, USP21, PEX10, B4GALT3, UBE2Z, DCTN5, LMAN2, MAN1B1, BTBD6,
   POMK, WDR5, H4-16, GPAA1, RAET1G, DYNLL2
   Unfolded Protein Response UPR 5.3394 0.0242 7 102 NFYA, ARFGAP1, WFS1,
   KLHDC3, TLN1, CTDSP2, SRPRA
   COPI-mediated anterograde transport 5.3922 0.0242 7 101 CAPZB, ARFGAP1,
   NAPA, GORASP1, TMEM115, DCTN5, DYNLL2
   Signaling by Receptor Tyrosine Kinases 2.5233 0.0242 18 555 AP2B1,
   ATP6AP1, COL5A3, PXN, POLR2C, MET, PPP2R5D, YAP1, DUSP6, PDPK1, GRB7,
   POLR2D, MAPKAPK2, YWHAB, STAT3, NELFB, PTPN1, COL5A2
   Integrin alphaIIb beta3 signaling 11.5262 0.0242 4 27 RAPGEF3, TLN1,
   PDPK1, PTPN1
   Integrin signaling 11.5262 0.0242 4 27 RAPGEF3, TLN1, PDPK1, PTPN1
   Platelet activation signaling and aggregation 3.1648 0.0283 12 295
   STXBP2, GNB1, RAPGEF3, GNA11, STX4, GNAI2, TLN1, PDPK1, CHID1, PTPN1,
   PSAP, ACTN4
   Activation of BH3-only proteins 9.7253 0.0409 4 32 TP53, YWHAB, TFDP1,
   DYNLL2
   Pausing and recovery of Tat-mediated HIV elongation 9.1532 0.0455 4 34
   POLR2C, CDK9, POLR2D, NELFB
   Tat-mediated HIV elongation arrest and recovery 9.1532 0.0455 4 34
   POLR2C, CDK9, POLR2D, NELFB
   Sphingolipid metabolism 5.2451 0.0455 6 89 CTSA, KDSR, CSNK1G2, SMPD1,
   CERS6, PSAP
   MET activates STAT3 38.9010 0.0455 2 4 MET, STAT3
   HIV elongation arrest and recovery 8.6447 0.0456 4 36 POLR2C, CDK9,
   POLR2D, NELFB
   Pausing and recovery of HIV elongation 8.6447 0.0456 4 36 POLR2C, CDK9,
   POLR2D, NELFB
   MHC class II antigen presentation 4.3569 0.0456 7 125 AP2B1, CTSA,
   CAPZB, AP1B1, ACTR1B, DCTN5, DYNLL2
   [132]Open in a new tab

   The pathway enrichment analysis for genes whose expression in the cone
   region of CE was found to be correlated with the eye rubbing manner
   (recognized by selecting one of the photographs presented in [133]Fig 1
   by study individuals) showed the enrichment in apoptosis,
   chaperon-related, and unfolded protein response pathways ([134]Fig 3).
   Particularly, the expression of TGFBI (R = 0.79, p = 0.0039), BCL2L1 (R
   = 0.65, p = 0.31), CAPZB (R = 0.73, p = 0.011), CTDSP2 (R = 0.74, p =
   0.0086), PTPN1 (R = 0.77, p = 0.0057), PDPK1 (R = 0.8, p = 0.0029),
   RAPGEF3 (R = 0.79, p = 0.0041), TFDP1 (R = 0.68, p = 0.022), SRPRA (R =
   0.72, p = 0.013), TLN1 (R = 0.62, p = 0.042), and TP53 (R = 0.78, p =
   0.0042) genes, was correlated with severity/potential harmfulness of
   the eye rubbing to the eyeball ([135]Fig 4).

Fig 4. Results of Pearson’s correlation.

   [136]Fig 4
   [137]Open in a new tab

   Pearson’s correlation results between the expression of selected genes,
   TGFBI, BCL2L1, CAPZB, CTDSP2, PTPN1, PDPK1, RAPGEF3, TFDP1, SRPRA,
   TLN1, and TP53 (in TPM), in the cone (central) region of corneal
   epithelium and severity/potential harmfulness of the eye rubbing to the
   eyeball recognized by selecting one of the photographs presented in
   [138]Fig 1 by study individuals. Values of regression and statistical
   significance are presented in plots.

   The expression of TRIM8 (R = 0.73, p = 0.0006), JUNB (R = 0.64, p =
   0.0043), and LTBP3 (R = 0.66, p = 0.0028) genes were found to be
   correlated with time spent using a computer after work, as presented in
   [139]Fig 5.

   Without taking into account the environmental/behavioral aspects in the
   correlation analysis, the expression of all the genes mentioned above
   did not differentiate the studied groups. ([140]S1 Fig).

Influence of allergy status on transcriptomic features and biological
interpretation

   Evaluating the allergic / non-allergic factors contributing to the
   phenotype we found that the genes whose expression was extrapolated to
   the allergy status didn’t contribute to IgE-related pathways, such as
   ‘allergen dependent IgE bound FCERI aggregation’. No allergy-related
   inflammation of any kind has occurred (e.g., ‘interleukin-4 and
   interleukin-13 signaling’), as presented in [141]S2 Fig, and [142]S7
   Table.

Discussion

   The genetic predisposition to KTCN, comprising the presence of sequence
   variation in genes related to KTCN [[143]26, [144]27] and KTCN-specific
   transcriptomic alterations [[145]19, [146]28] along with environmental
   [[147]9, [148]29], behavioral [[149]9, [150]30, [151]31], and other
   unknown factors, most likely cause the onset of KTCN.

   Previously the strong relation between eye rubbing and KTCN was
   reported in the case reports [[152]32], case series [[153]33],
   multivariate analyses [[154]30], and meta-analysis [[155]9]. Here we
   verified this aspect and further characterized the eye rubbing behavior
   practiced by patients with KTCN. Patients used their fists more
   frequently to rub the eye, and they also rubbed most of the eye
   surface, i.e., the upper and/or lower eyelid, as opposed to the inner
   and/or outer corner of the eye. Moreover, this disclosed manner
   involved the use of greater force when patients with KTCN rubbed their
   eyes. Previously, the mechanical forces exerted on the eyelids by KTCN
   patients by rubbing the eyes were reported to be different, depending
   on which part of the hand was used [[156]34].

   Comparing the corneal thickness and other clinical parameters between
   less and more frequently rubbed eyes, we didn’t find the substantial
   differences. Although we’ve revealed relation between eye rubbing and
   absolute differences in TCT (TCT diff) between right and left eye, the
   specific manner of eye rubbing (e.g., eye rubbing with fist) or its
   severity did not show correlation with this variable. Since central
   corneal thickness is one of the most heritable human characteristics
   [[157]35], we concluded that the effect of eye rubbing on the KTCN
   phenotype was rather due to the generation of cellular stress in the
   CE.

   Previously, in vitro study showed the alteration of cytoskeleton and
   migratory behavior of corneal epithelial cells exposed to shear stress
   [[158]36]. In line with these findings, here the increased expression
   of TFDP1, SRPRA and CAPZB, being elements of cellular responses to
   stress, and the increased expression of TLN1, RABGEF3, PDPK1, and
   PTPN1, being elements of extracellular matrix (ECM) affecting cell
   adhesion and migration were revealed, as the in vivo evidence of
   influence of severe eye rubbing on CE. Importantly, these results were
   consistent for the patients of both sexes, while previously reported
   transcriptomic data [[159]37] and epidemiological [[160]38]
   observations were sex-specific.

   Moreover, we revealed the important eye rubbing-related changes in
   expression of genes contributing to apoptosis (BCL2L1, TP53, BAX),
   chaperone activation (TLN1, CTDSP2, SRPRA), and unfolded protein
   response (NFYA, TLN1, CTDSP2, SRPRA). Previously, among the listed
   genes, high levels of protein p53 were found in unaffected CE of
   various vertebrate species [[161]39], and more importantly in the
   interpretation of our outcomes, the experimental study showed the
   increased cytoplasmic expression of p53 after the cell stress
   stimulation resulting from ultraviolet irradiation [[162]40].

   Notably, in a relation to severity of eye rubbing a dysregulated
   expression of elements of the TGFB signaling pathway was identified,
   previously reported as the KTCN-specific feature [[163]19] and
   discussed as influencing the activation of corneal cells to produce
   various types of collagen and other ECM components [[164]41].
   Upregulation of TGFBI revealed here was found to be corelated with more
   severe eye rubbing, whereas increased expression of JUNB and LTBP3 were
   strongly related to the hours spent in front of the computer after
   work. The longer time of using a computer after work was revealed as an
   independent risk factor for KTCN. However, patients with KTCN were
   younger and definitely more addictive to electronics and that should be
   considered in the data interpretation.

   Moreover, the presence of dust in the working environment increased the
   risk of both the eye rubbing and KTCN itself. Interestingly, this
   factor may have similar influence on corneal surface as the living in
   median size cities, in which exposure to air pollution, particularly to
   particles with size less than 2.5 μm, was reported to be higher as an
   effect of urbanization degree and household heating systems [[165]42,
   [166]43].

   Here, the allergy has not been disclosed as a risk factor for KTCN.
   Regarding the aspect of allergy and atopy in patients with KTCN, it has
   been previously discussed as more prevalent in KTCN than in general
   population [[167]30, [168]44, [169]45]. However, in some reports using
   multivariate analyses and meta-analyses, the atopy and allergic eye
   diseases were not significantly associated with KTCN [[170]46, [171]47]
   and in many developing countries the constant increase in occurrence of
   allergic and atopic diseases has been reported [[172]48, [173]49]. For
   example, the incidence of at least one allergic condition changed in
   years 2001–2005 from 18.9% to 24.2% in patients registered in England
   database (QRESEARCH 2,958,366 patients) [[174]50]. In the analysis of
   transcriptome in the full-thickness corneas no significance for the
   allergy status was found, no differentially expressed genes comparing
   the KTCN patients with allergy vs. without allergy were indicated
   [[175]51]. Moreover, no significant differences in the ocular surface
   immune cell subset proportions and tear fluid soluble factor levels
   were observed between KTCN eyes of different grades [[176]52]. Also in
   that study, the influence of ocular allergy, history of systemic
   allergy and eye rubbing on the immune cell profile was investigated,
   and no differences were found [[177]52]. In contrast, the significantly
   higher tear fluid levels of IL-1α, IL-9, IL-10, IL-13, TNFα, sVCAM,
   sTNFRII and IgE were observed in KTCN patients with history of systemic
   allergy and significantly higher levels of cytokines (IL-2,
   IL-12/23p40, IL-12p70, IL-13, IL-17A, IFNα), chemokines (MIG/CXCL9;
   ITAC/CXCL11), and growth factors (TGFβ1, EPO, VEGF) were revealed in
   KTCN patients with a history of eye rubbing [[178]52]. In our KTCN
   patients the allergy was found to be not related to the severity of eye
   rubbing. Importantly, allergy status has not affected CE transcriptomic
   outcomes obtained for the same evaluated patients. Also, other
   inflammatory conditions, reported in the past in medical history,
   conjunctivitis, both bacterial and viral and autoimmune diseases, were
   not found to be more frequently observed in patients with KTCN. We
   conclude here that the eye rubbing behavior is not a consequence of the
   allergic condition.

   We observed that patients with KTCN were more frequently left-handed
   comparing to control individuals being right-handed in general.
   Although we did not find a relation between the most frequently rubbed
   eye and right/left-handedness, the relation between the most frequently
   rubbed eye and the more affected eye, in particular between the left
   eye and the preference of left-eye rubbing, was revealed.

   It has been reported that patients with KTCN more frequently than
   controls manifested dry eye syndrome [[179]53]. Here the relation of
   KTCN with the dry eye syndrome, verified by TBUT and/or OSDI (data not
   shown), was not confirmed.

   As outdoor activity is a verified factor influencing other
   multifactorial eye diseases, the myopia and high myopia phenotypes
   [[180]54], and it has not been investigated in KTCN research, we also
   assessed this aspect in patients with KTCN as a risk or protective
   environmental factor and found it as not differentiating between the
   groups of patients and controls which we evaluated.

   Also, patients with KTCN less frequently read paper books and used the
   electronic readers, probably due to the joint effect of a generational
   change as well as a problem with visual acuity that cannot be adjusted
   through glasses or contact lenses.

   In this paper, we presented the level of higher education as the only
   variable reducing the risk of KTCN (OR = 0.52, univariate analysis),
   but this result should be treated with caution as it may be a
   consequence of selection bias.

   Other, previously discussed factors as sleeping position and creating
   additional pressure on the eyeball [[181]30] or smoking [[182]31]
   weren’t confirmed in our study. Moreover, the effect of pregnancy on
   KTCN progression could not be verified due to a small number of female
   patients in KTCN group. In this study more men than women were
   ascertained in KTCN group, which is in line with the predominance of
   male patients among 110 individuals undergoing the CXL procedure in
   2020–2022 in Optegra Clinic (19 female and 91 male patients,
   unpublished data, Maleszka-Kurpiel, MD, PhD, personal communication)
   and with the previous studies [[183]55, [184]56].

   As this is a single center patient recruitment study, we interpreted
   our results with caution. Questions asked, based on the applied
   Questionnaire, were somehow biased, for example in recalling of
   habits/behaviors and quantifying their degree. Therefore, in the case
   of the behavioral aspects, we used the multiple question method to
   minimize such biases. In addition, the study groups were not matched in
   terms of age and sex, as the aim of the study was also to analyze the
   impact of these characteristics. We did not identify clinical features
   corelated with sex or age, however, the sex and age could potentially
   influence study results regarding the life habits. The absence of
   differences in clinical parameters between adult and adolescent
   patients indicates a rapid progression in the latter [[185]57–[186]59],
   pointing to the importance of increasing patient awareness [[187]60] as
   well as early diagnosis due to substantial risk of progression.
   Generally, we focused on the regular patients, who best represent the
   overall patient population of the clinic. Hence in our control group,
   the KTCN family members occur, but differences in their life habits in
   comparison to KTCN-affected family members could be a valuable
   practical and clinical information. Importantly, the study groups do
   not differ in the aspect of contact lens wear, which could have
   influence the issue of frequency of the eye rubbing. Moreover, the
   absence of differences in clinical parameters between adult and
   adolescent patients, and male and female patients, in contrast to
   previous reports [[188]38, [189]61], enabled a superior appreciation of
   both, behavioral and molecular findings. Regarding other limitations,
   obtaining the CE samples from the control individuals towards
   transcriptomic assessment was especially challenging, as this type of
   material could be collected during PRK procedure only, which nowadays
   is rarely performed and replaced with the newer techniques of vision
   correction, in which the CE is not removed.

   So far, the impact of shear stress on function/mechanics of epithelial
   cells in single-layer model [[190]36] and changes in mechanical forces
   exerted on the eye depending on eye-rubbing manner [[191]34] have been
   reported. However, currently the effects of repetitive eye rubbing on
   corneal biomechanics have not been confirmed in a more complex model
   such as the ex-vivo porcine eyes [[192]62]. Therefore, the influence of
   the eye rubbing, confirmed in the corneal epithelium cultured cells
   under flow conditions [[193]36] should be further investigated in the
   studies on the biological mechanisms of this behavior on the physiology
   of the cornea, involving the 3D experimental models.

   Overall, environmental / behavioral factors were shown to have a
   significant influence on the corresponding KTCN CE transcriptomic
   findings in the patients with KTCN. More importantly, the evaluated
   transcriptomic differences found between KTCN and control samples were
   only revealed after taking environmental / behavioral factors into
   account.

   In conclusion, the male sex, eye rubbing, time of using a computer
   after work, and the dust in the working environment were found or
   verified as the most significant risk factors for KTCN and the effect
   of selected environmental/behavioral features on ocular surface in KTCN
   was revealed. Presented transcriptomic findings, integrated with
   environmental/ behavioral aspects, support the hypothesis of chronic
   mechanical corneal trauma affecting the pathogenesis of KTCN. This
   trauma in the form of eye rubbing-caused damage of the CE, leaving the
   allergy as a redundant condition, contributes to the cellular stress,
   which through influence on cell apoptosis, migration, and adhesion
   affects the KTCN phenotype.

Supporting information

   S1 Fig. The boxplots of selected genes’ expression.

   The boxplots presenting the expression of selected genes, TGFBI,
   BCL2L1, CAPZB, CTDSP2, PTPN1, PDPK1, RAPGEF3, TFDP1, SRPRA, TLN1, TP53,
   TRIM8, JUNB, and LTBP3 (in TPM, Transcripts per Million), in the cone
   (central) region of corneal epithelium in the studied groups of
   patients with KTCN and controls.

   (TIF)
   [194]Click here for additional data file.^ (725.9KB, tif)
   S2 Fig. Results of pathway enrichment analysis.

   Results of pathway enrichment analysis for genes whose expression in
   corneal epithelium was found to be correlated with the allergy status.
   Transcriptome data was obtained for males and females from both study
   groups. The Reactome database was chosen as the source of the pathways.

   (TIF)
   [195]Click here for additional data file.^ (767.4KB, tif)
   S1 Table. Results of clinical comparison of male and female patients
   with KTCN.

   (DOCX)
   [196]Click here for additional data file.^ (37.3KB, docx)
   S2 Table. Results of clinical comparison of adult and adolescent
   patients with KTCN.

   (DOCX)
   [197]Click here for additional data file.^ (49.7KB, docx)
   S3 Table. Complete information on the qualitative and quantitative
   data/results of analyzes of behavioral, environmental, and
   socioeconomic aspects.

   (DOCX)
   [198]Click here for additional data file.^ (49.2KB, docx)
   S4 Table. Results of comparison of male and female patients with KTCN
   in aspect of selected behavioral, environmental, and socioeconomic
   factors.

   (DOCX)
   [199]Click here for additional data file.^ (17.6KB, docx)
   S5 Table. Results of comparison of adult and adolescent patients with
   KTCN in aspect of selected behavioral, environmental, and socioeconomic
   factors.

   (DOCX)
   [200]Click here for additional data file.^ (17.5KB, docx)
   S6 Table. Additional statistical analyses.

   (DOCX)
   [201]Click here for additional data file.^ (53.2KB, docx)
   S7 Table. Results of pathway enrichment analysis for genes whose
   expression in corneal epithelium was found to be related to the allergy
   status.

   (DOCX)
   [202]Click here for additional data file.^ (13.9KB, docx)
   S8 Table. STROBE statement—checklist.

   (DOC)
   [203]Click here for additional data file.^ (90KB, doc)

Acknowledgments