Introduction

Keratoconus (KC) is a progressive ectatic disease with unknown pathogenesis, characterized by thinning and cone-like steepening of the cornea. The general prevalence of KC is 1/2,000, but recent studies using more advanced tomographic/topographic methods have reported rates of 1.5% to 3.6%.^1,2,3 KC often begins in adolescence and usually shows asymmetric involvement.^4,5,6 Although onset seems to occur in the second decade of life, a Netherlands-based study found that patients with KC were diagnosed at a later age (mean age: 28.3 years).⁷ Delayed diagnosis and difficulty in the early detection of progression affect the treatment approach algorithm in progressive KC.

Corneal cross-linking (CXL) therapy was developed to prevent progression of KC and also significantly reduces the need for keratoplasty.^8,9 The decision to perform CXL is based on monitoring of progression in KC patients over 18 years of age, whereas for pediatric patients the general approach is to perform CXL when the initial diagnosis is made because progression occurs in up to 88% in this group.^10,11

Although many different parameters are used to detect KC progression, there is no consensus on the definition of progression. Based on the ABCD KC staging system developed by Belin et al.¹², the Belin ABCD progression display was added to the Pentacam software in 2017, bringing a more systematic new approach to KC progression. This program presents the anterior (A) and posterior (B) radius of curvature in the 3-mm zone centered on the thinnest point of the cornea, the thinnest corneal thickness (TCT) (C), and best corrected visual acuity (D) values within the 80% and 95% confidence intervals of measurement variability in normal and KC eyes. Measurements beyond these variability confidence intervals are interpreted as indicating progression.

This study was conducted to examine the association between progression and baseline tomographic and biomechanical characteristics in KC patients with progression according to Belin ABCD analysis.

Materials and Methods

This retrospective study included 111 KC patients who were followed up in the corneal unit of the ophthalmology department of Eskişehir Osmangazi University Hospital between 2015 and 2019 and had at least 1 year of follow-up and 3 separate Scheimpflug corneal tomography (Pentacam HR, Oculus Optikgeräte GmbH, Wetzlar, Germany) measurements at intervals of at least 3 months and Ocular Response Analyzer (ORA, Reichert Inc., Depew, NY, USA) measurement at baseline. The study was conducted in accordance with the requirements of the Declaration of Helsinki after obtaining approval from the Eskişehir Osmangazi University Faculty of Medicine Non-invasive Clinical Research Ethics Committee (12.05.2020/07).

KC was diagnosed in the presence of slit-lamp findings such as Fleischer ring, Vogt striae, and apical scar; keratometry values (K1/K2) >48 diopters (D); and corneal tomographic findings consistent with KC such as maximum keratometry (Kmax) >49 D, axial distortion, inferior steepening, irregular astigmatism, abnormal posterior elevation, and abnormal corneal thickness distribution.¹ Patients with history of ocular surgery including CXL, penetrating keratoplasty, deep anterior lamellar keratoplasty, and cataract surgery, patients with corneal scarring and ocular surface problems, patients under 18 years of age, and patients with no potential for progression due to stage 4 (end-stage) KC according to topographic KC classification (TKC) were excluded.

Pentacam measurements performed at the patient’s first two consecutive visits at an interval of 3±1 months were evaluated separately on the Belin ABCD KC progression display. Progression was defined as any one of the parameters A, B, and C on this screen exceeding the ≥95% confidence interval for the KC patient population (solid red line) or any two of the parameters exceeding the ≥80% confidence interval for the KC patient population (dotted red line). We did not evaluate criterion D, visual acuity, because studies have shown it is not a valuable finding in terms of progression.^11,13 One eye of each patient was included in the study. Patients with progression in either eye were evaluated on a case basis as progression. In patients with unilateral progression, the progressive eye was included in the analysis; for patients with bilateral progression or no progression, the eye with better baseline values was included. As is routine practice in our clinic, Pentacam and ORA readings were performed at least 1 hour after removing contact lenses.

Pentacam parameters analyzed in relation to progression were Kmax, TCT, central corneal thickness (CCT), 90° vertical anterior and posterior coma, and Belin Ambrosio Enhanced Ectasia Display Final D value (BAD D). ORA parameters analyzed were corneal hysteresis (CH), corneal resistance factor (CRF) and applanation waveform parameters (p1area, p2area, uslope1, uslope2, dslope1, dslope2, w1, w2, h1, h2) (Table 1).

Statistical Analysis

IBM SPSS version 22.0 (IBM Corp, Armonk, NY, USA) was used for statistical analyses. Independent groups t-test and univariate and multivariate logistic regression analysis were used to evaluate factors associated with KC progression. Variables that showed significance in univariate logistic regression (p<0.05) and did not show multicollinearity were included in the multivariate model. Paired samples t-test was used to compare initial and final examinations within groups. A p value <0.05 was considered statistically significant.

Results

The mean age of the 111 patients included in the study was 29.4±9.0 years, 50 (45%) were male, and the frequency of progression within the mean follow-up period of 26.4±12.0 months was 39.6% (n=44). The male to female ratio in the progression and non-progression groups was 25/19 and 36/31, respectively, and the difference was not statistically significant (p=0.75). The progression group was younger than the non-progression group (mean age 27.1±8.5 and 31.1±9.1, respectively, p=0.02). The mean follow-up time was longer in the progression group (29.5±10.9 and 24.4±11.7 months, respectively, p=0.03). During follow-up, there was a statistically significant increase in Kmax and mean keratometry (Kmean) values (p<0.05) and marginally significant decreases in CCT and TCT in the progression group (Figure 1). These four parameters were stable in the non-progression group during follow-up (Figure 1).

The TKC stages at initial examination of patients in the progression and non-progression groups are shown in Table 2. According to TKC staging, 63.6% (28/44) of patients with progression and 58.2% (39/67) of those without progression were stage 2 or 3. The ABC criteria indicating progression in patients in the progression group were, in order of frequency, B (84%), A (77%), and C (64%) (Figure 2).

The two groups had similar initial mean values for Kmax, Kmean, TCT, CCT, 90 vertical anterior and posterior coma, final BAD D, minimum/maximum/mean Ambrosio-related thickness (ART Min/Max/Avg, respectively) (p>0.05 for all; Table 3). Although the mean initial CRF and CH values were also similar in both groups (p>0.05), the p2area, uslope2, dslope1, h1, and h2 values obtained from the waveform were significantly lower in the progression group compared to the non-progression group (p=0.026, 0.036, 0.021, 0.034, and 0.029, respectively; Table 3).

In the univariate logistic regression analysis examining tomographic and biomechanical factors associated with KC progression with correction for age, sex, Kmax, and follow-up time, none of the initial Pentacam variables were associated with progression (Table 4). However, progression was associated with the ORA parameters h2 (for each 10-unit increase, odds ratio [OR]: 1.06, 95% CI: 1.01-1.11, p=0.03) and p2area (for each 100-unit increase, OR: 1.08, 95% CI: 1.01-1.15, p=0.02; Table 4) derived from the second applanation wave. In the same analysis, age (for each year increase, OR: 1.07, p=0.01) and follow-up period (for each year increase, OR: 1.60, p=0.03; Table 4) were also associated with progression. In the multivariate model, age, follow-up time, and p2area were found to be independent determinants of progression (Table 4). Superimposition of the ORA applanation curves of patients with and without progression showed that patients with progression had a relatively earlier applanation in the first applanation and later recovery in the second applanation, and lower height in both the first and second applanation curves compared to patients without progression (Figure 3).

Discussion

In this study, progression defined according to parameters A, B, and C on the Belin ABCD progression display was observed in 39.6% (n=44) of the patients. In patients with progression, we observed that Kmax and Kmean increased by 1.0 D and 0.5 D, respectively, and CCT and TCT decreased by approximately 5 µm during follow-up (Figure 1). The parameter most effective in determining progression was posterior surface radius of curvature (criterion B, 84%), followed by anterior surface radius of curvature (criterion A, 77%) and thinnest pachymetry value (criterion C, 64%) (Figure 2). KC progression was associated with younger age (for each additional year, OR: 1.08, p=0.006), longer follow-up time (for each additional year, OR: 1.78, p=0.01), and lower p2area on initial ORA measurement (for each 100-unit increase, OR: 1.07, p=0.01) (Table 3).

According to the 2015 global consensus report on KC and ectatic diseases created by Delphi panel, progression was defined as meeting at least two of three criteria (steepening of the anterior corneal surface, steepening of the posterior corneal surface, and corneal thinning and/or an increase in the rate of thickness change from the peripheral cornea to the thinnest point), but it was not clearly stated what amount of change in these parameters should be considered progression.¹¹ Many topographic/tomographic parameters are used in routine progression monitoring. The most important of these, Kmax, represents only the anterior surface of the cornea but does not provide information about the posterior surface, may vary in patients using hard gas-permeable contact lenses, and has been reported to remain unchanged or even decrease in progression, resulting in controversy regarding its use in the follow-up of progression and CXL effectiveness.^14,15,16,17 This idea is supported by our finding that the radius of curvature of the posterior corneal surface (criterion B) was a more frequent sign of progression than the anterior surface radius of curvature radius (criterion A) (84% and 77%, respectively, Figure 2). In a meta-analysis examining changes in other parameters used in progression monitoring during the natural course of KC, it was reported that best corrected visual acuity and sphere/cylinder values did not show statistically significant changes during the follow-up period and thus its use in follow-up would not provide meaningful results.¹⁸

The Belin ABCD progression display offers a different perspective on progression based on changes in the ABCD staging system developed by Belin et al.¹² Unlike the Amsler-Krumeich classification, it also takes into account the posterior corneal surface, evaluates not the entire cornea but the central cone where the main changes are seen (3-mm area centered on the thinnest point of the cornea), presents separate ratings on the basis of 4 parameters, and is consistent with the criteria proposed in the global consensus report.^15,16,17 Kösekahya et al.¹⁹ compared the Belin ABCD progression display with traditional criteria (Kmax, CCT, and anterior/posterior elevation changes) in the detection of progression and showed that the Belin ABCD progression display could provide an acceptable level of differentiation.

In the literature, parameters reported to have predictive value in KC progression include young age,^20,21,22,23 low TCT,²⁴ high Kmean,²⁴ high anterior Kmax^18,23 and posterior Kmax,²⁵ high central posterior²⁴ or anterior²⁰ elevation, index of surface variance (ISV),²⁶ high index of height decentration (IHD),²⁶ and vertical coma.²⁵ The fact that none of the initial corneal tomography parameters had predictive value for progression in this study whereas some initial biomechanical parameters showed significant differences between the groups supports the argument that the primary derangement in KC is biomechanical and that tomographic changes follow biomechanical disruption.^27,28,29 Age is a confirmed surrogate for corneal biomechanics³⁰ and was a significant predictive of progression in previous studies as well as our own, which strengthens the likelihood that biomechanical changes in KC are more important in the early stage.

The non-significance of CRF and CH, the basic ORA parameters, in our study while some waveform parameters showed significance requires explanation. CRF and CH are calculated based on the pressure difference between the first and second applanations, and CRF is calculated using a coefficient that emphasizes the first applanation. Although these two parameters are lower in eyes with KC, it has been reported that they have low diagnostic sensitivity and specificity and their ranges may overlap in normal and KC eyes.³¹ Although the importance of the ORA waveform parameters is still not clearly understood, various assumptions have been put forward.³² p1area and p2area are proportional to the time required for the cornea to transition from its natural convex shape to concave and back again; w1 and w2 (applanation width) are proportional to the transition speed of the cornea between convex and concave forms, h1 and h2 (applanation height) are proportional to the amount of light reflected from the corneal surface to the detector during applanation, and lower values for these parameters are suggested to be associated with a weaker corneal structure.^33,34,35 In parallel with the findings of our study, the presence of other studies demonstrating that waveform parameters have greater diagnostic value than pressure-based parameters (CRF and CH) in eyes with early-stage KC suggests that these parameters may be better biomechanical indicators.^36,37 In another study evaluating biomechanical changes with ORA before and after CXL, p2area, which is the most valuable predictive parameter in progression, was reported to be the parameter that best demonstrates biomechanical changes after CXL, but there was no significant change in CRF and CH values.³⁸ In a study by Küçümen et al.³⁹ examining changes in CRF and CH after CXL, no statistically significant change was observed in CH in the early or late postoperative period, while the change in CRF showed early significance that disappeared in the late postoperative period.

In our study, when ORA curves were averaged for all eyes with and without progression, we observed that corneas showing progression flattened earlier, started to recover later, and had significantly lower wave height for both applanations. Earlier applanation of a biomechanically weaker cornea is a finding that can be explained biologically. On the other hand, the air puff continues for a short while after the first applanation and the cornea becomes concave, and the transition back to the cornea’s normal state seems to be prolonged in progressive eyes. This may also be due to the relationship between the maximum concave radius of curvature, which is also a parameter of Corvis ST, and biomechanical strength. This radius is more resistant to deformation and has higher values in biomechanically stiff eyes.³¹ In other words, a weak cornea forms a deeper concavity when subjected to the air puff and thus takes longer to normalize, while a stronger cornea forms a shallower concavity and has a shorter normalization.

Study Limitations

One of the limitations of this study may be the exclusion of Belin ABCD progression criterion D (visual acuity) from our evaluation. However, although progression has been associated with a decrease in visual acuity in many articles in the literature, the widespread view in recent years is that uncorrected and best corrected visual acuity are not significant criteria for demonstrating progression.^11,18 Evaluating progression according to a change based on a single initial visit would increase variability and thus the false positivity rate; therefore, it has been stipulated that change based on two consecutive initial visits must be seen to be called progression. Since patients under the age of 18 were excluded from this study, our findings may not be valid for pediatric cases. Although the progression group had longer follow-up, statistical correction was made for the follow-up period in our analysis of predictive factors. Some patients included in the study used hard gas permeable or soft contact lenses. Although corneal contact lenses vary according to fitting choice, they may cause changes in the curvature of the anterior and posterior surfaces due to the mechanical effect and the hypoxia they cause, and even if the contact lenses are removed, the stabilization process may take several weeks.^40,41 Because it is not possible to wait this long in practice, measurements were obtained from our patients at least 1 hour after lens removal, as per routine practice in our clinic. The extent to which the ORA device performs a true biomechanical assessment is controversial, as the pressure and waveform-based parameters obtained with the ORA are seriously affected by the geometric properties of the cornea (e.g., thickness) and by intraocular pressure. In addition, the ORA has variable air puff pressure and utilizes an infrared camera that can provide low-resolution data, which are shortcomings compared to the Corvis ST.³¹ For this reason, conducting similar studies with new generation devices such as Corvis ST and Brillouin spectroscopy, which provide advanced biomechanical evaluation, will better elucidate the extent to which biomechanical properties are associated with progression.

Conclusion

In conclusion, this study demonstrates that younger age and biomechanical properties may be indicators of future progression and that tomographic parameters follow biomechanical changes. This study is the first report showing that the biomechanical parameters obtained with the ORA device may be important in predicting the progression of KC.

Ethics

Ethics Committee Approval: Ethics committee approval was obtained from Eskişehir Osmangazi University Non-Interventional Clinical Research Ethics Committee with the decision no. 07 dated 12.05.2020.

Informed Consent: Retrospective study.

Peer-review: Externally peer reviewed.

Authorship Contributions

Concept: E.A., M.A.E., Design: E.A., M.A.E., N.Y., Data Collection or Processing: M.A.E., O.Ö., A.D., Analysis or Interpretation: E.A., M.A.E., N.Y., Literature Search: E.A., M.A.E., N.Y., A.D., Writing: E.A., M.A.E., O.Ö.

Conflict of Interest: No conflict of interest was declared by the authors.

Financial Disclosure: The authors declared that this study received no financial support.