Corneal biomechanical properties in vernal keratoconjunctivitis and its subtypes: a preliminary study

To explore the corneal biomechanical properties (CBPs) of patients with vernal keratoconjunctivitis (VKC) and the discrepancies among three subtypes of VKC including palpebral, limbal, and mixed forms. Forty eyes of 20 VKC patients and twenty eyes of ten non-VKC patients were included in this case–control study. Patients with VKC were further divided into three subtypes (six patients in Palpebral form, five patients in limbal form, and nine patients in mixed form). The CBPs of all patients were obtained from the Corneal Visualization Scheimpflug Technology (Corvis ST). First applanation (A1) length, Ambrosio relational thickness in horizontal (ARTh), and stiffness parameter at first applanation (SP-A1) were significantly lower in the VKC group while A1 velocity was significantly higher in the VKC group (p < 0.05), compared to the non-VKC group. Furthermore, A1 velocity presented a positive correlation with disease course (p < 0.05). In addition, VKC patients of limbal form had lower central corneal thickness (CCT), SP-A1, and higher deformation amplitude ratio (DA ratio), compared to the other two subtypes (p < 0.05). Besides, patients in limbal form had higher A1 velocity, integrated radius, and corneal biomechanical index (CBI) compared with mixed form, and lower A1 length than palpebral form (p < 0.05). The corneas of VKC patients were softer and more protruded compared with the control group, and the property of steepness was closely related to disease course. VKC patients in limbal form were more inclined to be keratoconus than the other two subtypes due to their CBPs` discrepancies.


Introduction
Vernal keratoconjunctivitis (VKC) is a chronic, bilateral, allergic disease of the eye and characterized by conjunctival congestion, itching, and ropy discharge [1]. The prevalence of VKC is 3.3-5.8% and children are the main crowd of VKC [2,3]. A randomized controlled trial demonstrated that up to 11.1% of children are suffering from VKC and the risk factors of VKC include male as well as personal systemic allergy history [4]. VKC is consisted of three subtypes, including palpebral form, limbal form, and mixed form. The palpebral form in VKC is featured as many giant papillae located in the upper tarsal conjunctiva, which looks like cobblestone. The signs of limbal form in VKC include gelatinous infiltrates of the limbus, corneal epithelium injury, shield ulcer, and so on. The mixed form means the physical signs are a combination of both forms [5]. Keratoconus (KC), a common complication of VKC, is characterized by progressively thinning of the corneal stroma, increasing asymmetric corneal curvature, and visionthreatening [6]. Among VKC patients, the prevalence of KC could be as high as 26.8% and 71% of them appear changes in corneal biomechanics [7].
Corneal biomechanical properties (CBPs) play a vital role in the structure and functions of the cornea. For instance, CBPs are the foundation of corneal shape and further determine the refractive function of the cornea. CBPs such as corneal viscoelasticity and corneal stability can also affect the correction of ametropia [8]. Moreover, central cornea thickness (CCT) and intraocular pressure (IOP) are closely related to CBPs. CCT is decided by the extent of the corneal extension under the load of IOP, which can be influenced by CBPs. A lower stiffness of the cornea leads to a lower CCT and IOP, while both CCT and IOP are significant indicators in ophthalmological examination or pre-surgical evaluation [9].
Nowadays, researches between VKC and CBPs are limited. Although Emre et al. [10] demonstrated that the corneal resistance factor and the Goldmann-correlated IOP were significantly lower in VKC patients, other CBPs changes in patients with VKC are still lacking. It is also unknown whether three subtypes of VKC have different effects on corneal biomechanics. Therefore, this study aims to further explore the changes in CBPs and the discrepancies among three subtypes in VKC patients.

Participants
This study was conducted according to the Declaration of Helsinki and informed consents were obtained from the parents of all participants. Forty eyes of 20 VKC patients and twenty eyes of ten non-VKC patients from the Eye Ear Nose and Throat Hospital of Fudan University were included in this case-control study. The inclusion criteria of the VKC group contained clinical features (itching, redness, sense of foreign body, mucous discharge, and seasonal recurrence) and signs (congestion of tarsal or bulbar conjunctiva, hyperplasia of papillary or follicular, and gelatinous infiltrates of limbus). In accordance with various signs, patients in the VKC group were divided into three subtypes (twelve eyes of six patients in Palpebral form, ten eyes of five patients in limbal form, and eighteen eyes of nine patients in mixed form). Subjects from the non-VKC group were minor ametropia and were void of other ophthalmologic or systematic diseases. No participants of the two groups had a history of contact lens wearing or ocular surgery, and the corneal epitheliums of all participants were intact.

Measurement of corneal biomechanical properties
Parameters of corneal biomechanics were obtained from the Corneal Visualization Scheimpflug Technology (Corvis ST). All participants were required to stare at the red spot in the instrument, and the Corvis ST would be automatically activated to conduct a measurement when it identified the correct position of the eyes. After that a signal containing the values of CCT, IOP, biomechanically-corrected IOP (bIOP), first applanation (A1) length, A1 velocity, second applanation (A2) length, A2 velocity, integrated radius, deformation amplitude ratio (DA ratio), Ambrosio relational thickness in horizontal (ARTh), stiffness parameter at first applanation (SP-A1) and Corvis biomechanical index (CBI) was sent to the computer attached to the Corvis ST. A trained ophthalmologist was arranged to perform this ocular examination.

Statistical analysis
Statistical analysis was conducted with Stata version 15.1. Means ± standard deviation (SD) were used to report continuous variables data. The normality of continuous variables was determined by the Shapiro-Wilk normality test. For the comparisons between the VKC group and the non-VKC group, the independent sample t test was used to evaluate the differences between the two groups, whose values presented normally distributed and passed the homogeneity test for variance. The Wilcoxon rank-sum test was used when values were not normally distributed. If values were categorical variables while a total number of that was under 40, the Fisher`s exact test was applied.
To judge the relational degree between the course of the disease and other CBPs, Pearson correlation analysis was used if data followed a bivariate normal distribution. If not, Spearman rank correlation analysis was applied. Correlation coefficient R > 0.3 or Rho > 0.3, R > 0.5 or Rho > 0.5, R > 0.8, or Rho > 0.8 was evaluated as mildly, moderately, and strongly correlated, respectively. After that linear regression was performed to present their quantitative interdependence relationship.
Concerning the contrasts among three subtypes in the VKC group, one-way analysis of variance (ANOVA) and Bonferroni test were performed if values were normally distributed and equal variance, while Kruskal-Wallis equality-of-populations rank test and Scheffe test were used to analyze non-normal distribution parameters. A value of p < 0.05 was considered statistically significant.

Characteristics of participants
Forty eyes of 20 VKC patients and twenty eyes of ten non-VKC patients were enrolled between April 2021 and February 2022. Sample demography is shown in Table 1. There was no significant difference in the age and sex ratio of the two groups (p > 0.05 for all). Additionally, clinical detailed features of 20 VKC patients are listed in Table S1. 3/20 of included participants were treated with steroid plus immunosuppressive eyedrops. 6/20 and 7/20 of VKC patients were applied with immunosuppressive and antihistamine eyedrops separately. Additionally, 4/20 of VKC patients had not taken any medication.
Eighteen VKC patients with accurate disease courses were involved to make the correlation analysis between the course of disease and CBPs, and its outcomes are listed in Table 3. A1 velocity had a moderate positive correlation with disease course  Fig. 1 and the equation of regression was y = 0.0034543x + 0.147592 (p = 0.03). Due to the limitation of sample numbers, no statistical difference had been found in the correlation analysis between other CBPs and the disease course so far (p > 0.05).
Comparisons of corneal biomechanical features among three subtypes in the VKC group are presented in Table 4. Statistical analysis demonstrated significant differences for CCT, A1 length, A1 velocity, DA ratio, integrated radius, SP-A1, and CBI (p < 0.05). No statistical difference was found in IOP, bIOP, A2 length, A2 velocity, and ARTh (p ≥ 0.05).
The further comparisons between any two subtypes in the VKC group are shown in Fig. 2. Patients of the limbal form had lower CCT and SP-A1, higher DA ratio and CBI, compared to the other two subtypes (p < 0.05) (Fig. 2a-c). A1 velocity and integrated radius were higher in limbal form than those in palpebral form, while A1 length of limbal form was lower, compared to mixed form (p < 0.05) (Fig. 2d-g). Nevertheless, there was no statistical difference in IOP, bIOP, ARTh, A2   length, and A2 velocity in comparisons between any two subtypes of the VKC group (Fig. 2h-l).

Discussion
VKC is a chronic allergic disease of the eyes and occurs mainly in boys, which is consistent with the demography of participants in our study. Children with VKC confront risks from ocular surface remodeling that may lead to eye-sight loss. A clinical study showed that mixed type VKC patients had lower corneal resistance factor and Goldmann-correlated IOP, while researches about three subtypes of VKC patients and changes in their CBPs are still deficient [10]. Therefore, we further explored the CBPs changes between VKC patients and non-VKC patients, and first carried out a preliminary study to distinguish the CBPs discrepancies among three subtypes in VKC patients.
In 2013, Emre et al. [10] used ocular response analyzer (ORA) to detect corneal biomechanical changes between mixed form VKC patients and healthy control subjects with minor reflective errors. The authors found that the measurements of corneal resistance factor and Goldmann-correlated IOP were lower in VKC patients (9.5 ± 1.7 versus 10.8 ± 1.7, p = 0.016; 13.3 ± 3.4 versus 16.6 ± 3.6, p = 0.013). The corneal resistance factor is the overall indicator of corneal resistance and is dominated by corneal elastic properties. They thought the decrease of corneal resistance factor was related to the impairment of linkages between keratocytes or collagen bundles, which might be caused by the abnormal immunological environment of cornea, conjunctiva, and tear film in VKC patients. Although Emre et al. [10] discovered mixed type VKC patients had significantly lower Goldmann-correlated IOP, we did not find any difference in either IOP or bIOP between VKC patients and non-VKC patients. In addition to it, Wang et al. [11] also found no significant difference in IOP or bIOP between allergic conjunctivitis patients and non-allergic conjunctivitis patients.
In our study, VKC patients had lower A1 length, ARTh, SP-A1, and higher A1 velocity compared to non-VKC patients. A1 length is the length of the flattened cornea at first applanation and A1 velocity is the velocity of the corneal apex during the first applanation [12]. The lower A1 length and higher A1 velocity indicated the shape of the cornea was more protruding, which was in accord with the results of ARTh. ARTh is the Ambrosio relational thickness in the horizontal 8 mm cornea, and the lower values of it mean that the measured cornea is steeper in shape [13]. In this study, ARTh in the VKC group was 452 ± 121.2 while ARTh in the non-VKC group was 518 ± 101.2 (p = 0.007), showing an obvious decline. Moreover, SP-A1, a stiffness parameter at first applanation, presented a significant decrease in our study as well (99.8 ± 17.7 in the VKC group versus 110.2 ± 17.2 in the non-VKC group, p = 0.04), indicating corneas of VKC patients were softer. ARTh, SPA1, A1 velocity, DA ratio, and integrated radius are proved to be stable individual parameters in CBI, which provides proof of keratoconus (KC) propensity [14]. In combination with the above findings, we postulated that the steeper shape and softer texture of corneas might cause VKC patients to tend to become KC patients.
Furthermore, we discovered that the course of the disease was related to values of A1 velocity. With the progression of the disease, A1 velocity became higher, which indicated that the corneas of VKC patients would become steeper in shape gradually. The pathogenesis of VKC is still unclear, immunity and inflammation were supposed to drive disease progress [15]. Kumagai et al. [16] found that corneal stromal fibroblasts could be activated by some bioactive substances in tear fluid such as proinflammatory and T helper 2 cell-derived cytokines, due to the incomplete corneal epithelium in VKC patients. Then, the activated fibroblasts of the corneal stroma enhanced the allergic reaction of the ocular surface by the means of promoting the expression of some chemokines and adhesion molecules and the infiltration of eosinophils and Th2 lymphocytes. As the result of chronic inflammation and immune reaction, further consequences of corneas in VKC patients were corneal stem cell deficiency, tissue remodeling reactions, superficial corneal opacification and so on [17].
We first conducted comparisons about the CBPs of three subtypes in VKC patients. The CCT of limbal form in VKC patients was significantly lower than the other two forms (514 ± 11 μm in limbal form; 548 ± 20 μm in palpebral form; 565 ± 23 μm in mixed form; p < 0.001), whereas no significant difference was found between palpebral form and mixed form.
Similar to CCT, SP-A1 was also lower in limbal form compared to the other two forms (79.9 ± 9.4 in limbal form; 100.3 ± 15.4 in palpebral form; 110.5 ± 13.0 in mixed form; p < 0.001), illustrating the patients with limbal form VKC had a softer cornea. In keeping with the outcomes of SP-A1, the measurement of the integrated radius was significantly higher in limbal form, which means the lower corneal stiffness of patients with limbal form VKC. Moreover, DA ratio as an index for delineating the corneal deformability was significantly higher in limbal form than in the other two forms as well (5.39 ± 1.7 in limbal form; 4.36 ± 0.3 in palpebral form; 4.24 ± 0.3 in mixed form; p < 0.001). The above outcomes showed the CBPs of limbal form were thinner, softer, and easier to deform, whereas the mechanism for that was unclear. In addition, the A1 length of limbal form was lower than that of palpebral form while A1velocity and integrated radius were higher than mixed form, demonstrating the corneal shape of limbal form was more protruding and its cornea was softer as well. Last but not least, CBI is a reliable diagnostic index for KC with a cutoff value of 0.5 [18]. In this study, we also found that values of CBI were significantly higher in limbal form (0.49 ± 0.35 in limbal form; 0.12 ± 0.19 in palpebral form; 0.13 ± 0.18 in mixed form; p = 0.02), demonstrating VKC patients of limbal form might be more possible to get KC. Totan et al. [19] studied the incidence of KC in VKC patients, they found that 26.3% of the mixed form subjects (10 of 38 subjects), 22.2% of palpebral form subjects (8 of 36 subjects), and 50.0% of the limbal form subjects (4 of 8 subjects) were KC patients, which also confirmed that VKC patients in limbal form were more likely to evolve into KC patients. Although VKC patients frequently rubbed their eyes due to severe itching of eyes, Totan et al. [19] did not find the chronic problem in most subjects of their research so they supposed hereditary or environmental factors were the main causes to propel VKC patients into getting KC. Combined with our study, corneas of patients with limbal form VKC showed differences in CBPs compared to the other two subtypes though subjects from all three subjects had a history of eye rubbing, which also indicated factors such as ocular immunity and heredity were the potential pathogeny rather than mechanical rubs.
In consideration of the side effects of steroid, general ophthalmologists tended to choose antihistamine or immunosuppressive eyedrops before giving patients a referral to corneal disease faculty. Therefore, only three patients in our research used steroid eyedrops. We noticed that Yu et al. [20] found the stiffness of rabbit cornea would get a 33.5% reduction after applied with steroid eyedrops for two months. Whereas, the effects that immune-suppressive or antihistamine played on CBPs had not been reported so far. On account of the small number of cases using steroid eyedrops, the above conclusion could not be verified in this paper.
The limitation of this primary study was the small sample size. Only 20 VKC patients participated in this study. Despite the subtype analysis, the numbers of the three subtypes were uneven. A prospective clinical trial with a larger sample scale should be designed to further verify the difference between VKC patients and non-VKC patients and the disparity among three subtypes in the VKC group.

Conclusion
The morbidity crowd of VKC is children, who are in the state of vision development [21]. The discoveries of this study demonstrated that the corneal biomechanics of VKC patients had been changed. The corneas of VKC patients were proved to be softer and steeper. In addition, with the prolongation of the disease course, the corneal shapes of VKC patients tended to be more protruding in shape. Furthermore, patients with limbal form VKC should be more cautious about the changes of their CBPs on account of the thinner, softer, more deformable cornea and a higher possibility of being KC compared to the palpebral form and the mixed form. Termly examination of corneal biomechanics was recommended for all VKC patients. Once finding the subclinical KC signs from CBPs changes, timely treatment such as anti-allergic drugs, steroids drugs, immunosuppressive agents, and corneal collagen crosslinking should be taken in case impairing the eyesight [22,23].