Corneal Biomechanics for Ocular Hypertension, Primary Open-Angle Glaucoma, and Amyloidotic Glaucoma: A Comparative Study by Corvis ST

Introduction

Corneal biomechanics has gained particular interest when the Ocular Response Analyzer (ORA; Reichert Ophthalmic Instruments, Depew, New York) was launched in the early 20^th century.¹ Later, Corvis ST (Corneal Visualization Scheimpflug Technology, Oculus; Wetzlar, Germany) was introduced. Both devices are noncontact tonometers that use an air puff to applanate the cornea. Corvis ST performs a comprehensive and detailed evaluation of the cornea through cross-sectional images captured by a high-speed Scheimpflug camera during the deformation cycle, providing more data than ORA.

In glaucoma, corneal biomechanics may contribute to elucidate the pathophysiologic mechanisms leading to axonal degeneration that are still unclear and to optimize the accuracy of intraocular pressure (IOP), the only proved modifiable risk factor for glaucoma onset and progression.²

It has been speculated that biomechanical properties of peripapillary sclera influence the damage to the optic nerve head (ONH) in response to variations of IOP.³ An excessive pressure-induced deformation of the lamina cribrosa (LC) eventually triggers the dysfunction and death of the retinal ganglion cell axons, and it may be exacerbated when the peripapillary sclera is more deformable. Currently, it is challenging to assess the biomechanical behavior of the posterior sclera and LC in vivo. The cornea shares a similar constitution on collagen fibrils and proteoglycan-rich extrafibrillar matrix with sclera.⁴ Therefore, corneal biomechanics has been regarded as an indirect measure of scleral and LC elasticity, and some studies have found correlations between corneal biomechanical parameters and the indexes of glaucoma progression.^5,6 The biomechanical properties of the cornea have been investigated with the aim of helping to characterize the individual susceptibility to glaucomatous damage.

Currently, all medical and surgical treatments aim to lower IOP. Several types of tonometers emerged to minimize the influence of corneal factors on IOP, but Goldmann applanation tonometry (GAT) remains the gold standard. GAT estimates the amount of force that is required to flatten an area of the cornea with a diameter of 3.06 mm.⁷ As such, the accuracy of IOP measured by GAT (GAT-IOP) is affected by corneal characteristics. Central corneal thickness (CCT) was the first parameter shown to influence the IOP measurement.⁸ Nowadays, beyond pachymetry, other corneal biomechanical parameters can be measured using ORA or Corvis ST and an IOP corrected for corneal biomechanics is further provided by both devices. To date, it has not yet been elucidated whether the biomechanically corrected IOP (bIOP) is more accurate than GAT-IOP and which is the role of bIOP in clinical practice.

In this study, we evaluated the biomechanical parameters of the cornea provided by Corvis ST in patients with primary open-angle glaucoma (POAG), amyloidotic glaucoma (AG), and ocular hypertension (OHT), and compared with healthy controls.

Methods

This was an observational cross-sectional study. Patients were recruited between July and September 2020 from ophthalmology outpatient clinic of a tertiary center, Centro Hospitalar Universitário do Porto. The research adhered to the principles of the Declaration of Helsinki and its latest amendment (Brazil, 2013), and complied with the requirements of the institute’s committee on human research.

To be included in this study, the diagnosis made by a glaucoma senior specialist (A.FIG, I.S., R.R. and M.J.M.) had to match one of the group’s definition. Healthy controls had an untreated IOP lower than 21 mmHg, normal optic discs in fundoscopy, and no other ocular disorders (except for senile cataract). POAG and OHT were defined according to European Glaucoma Guidelines, 4^th edition.⁸ AG was classified as a secondary open-angle glaucoma affecting patients with the genetic diagnosis of hereditary transthyretin-related amyloidosis (ATTR), and amyloid deposition along the pupil border in the slit-lamp examination.

The exclusion criteria included age below 40 years, high refractive errors (spherical refraction >6 diopters and cylinder correction >3 diopters), corneal disease, contact lens wear, shallow anterior chamber (Van Herick grading <3), uncontrolled GAT-IOP (>21 mmHg), previous glaucoma or refractive surgery, any ophthalmic surgery in the last 6 months, and a quality score (QS) other than “OK” in Corvis ST.

A comprehensive ophthalmologic examination was performed including a review of medical history, best-corrected visual acuity (BCVA) evaluation, slit-lamp biomicroscopy, GAT, fundoscopy, and Corvis ST, in the same schedule. GAT-IOP was corrected for CCT (GAT-IOPadj) based on the validated Ehler’s correction algorithm.⁹ BCVA registered in decimals was converted to the logarithm of the minimum angle of resolution (LogMAR) equivalent.¹⁰

Corvis ST (software version 1.6r2015) is a noncontact tonometer coupled with a high-speed Scheimpflug-camera (4330 frames/sec) to record the movements of the cornea in response to an air puff, which are then displayed on the built-in control panel in ultra-slow motion. The air puff forces the cornea through distinct phases: an ingoing phase in which the cornea passes from its resting shape through a first applanation (A1) into a concave shape (highest concavity, HC); an outgoing phase, which features a second point of applanation (A2) prior to the cornea returning to its normal resting state. Several biomechanical parameters are recorded during the deformation cycle. For the three main points (A1, HC, and A2), Corvis ST calculates the time, velocity, deformation and deflection amplitudes, and area and length of deflection. At HC, peak distance (PD) and radius (Rad) are additionally measured. Mathematically derived parameters, such as deformation amplitude ratio (DA ratio), integrated radius (1/R), stiffness parameter at A1 (SPA1), and stress–strain index (SSI), were recently integrated into the data output of Corvis ST. A more deformable cornea is characterized by lower time (A1T), smaller deflection length (A1DeflL) and higher velocity (A1V), at A1; higher deformation (HCDA) and deflection amplitudes (HCDeflA), and smaller PD and Rad, at HC; smaller deflection length (A2DeflL) and lower velocity (A2V), at A2; higher DA ratio and 1/R; lower SPA1 and SSI.^11–13 The opposite is found in stiffer and less deformable corneas. Furthermore, bIOP and CCT (with good reproducibility results)^14,15 are provided by the device.

Primary outcome was the comparison of corneal biomechanical parameters provided by Corvis ST between study groups. Secondary outcome was the comparison of different IOP measurements, particularly GAT-IOP, GAT-IOPadj, and bIOP, in each group.

All statistical analyses were performed using IBM SPSS^® software, version 22.0 (SPSS, Inc, Chicago, IL). Shapiro–Wilk test, Kolmogorov–Smirnov test and normal probability plots were used to confirm the normal distribution of the data. Differences among the study groups were evaluated with the use of an analysis of variance (ANOVA) model, followed by the post hoc Bonferroni test when the findings with the ANOVA model were significant and equal variances assumed. When equal variances were not assumed, the significance was tested with Welch’s t-test, followed by the post hoc Games–Howell test. For nonparametric variables, Kruskal–Wallis test was conducted and significant values were adjusted with the Bonferroni correction, using Dunn post-test. A linear mixed model was designed to assess the status of Corvis ST parameters between different groups adjusting for confounding variables (age, gender, GAT-IOP, and prostaglandin analogues medication). The patient identification number was included as a random effect to correct for the inclusion of both eyes in some participants. Statistical significance was defined as p <0.05.

Results

This study included 183 eyes of 115 patients: 61 with POAG, 32 with AG, 37 with OHT and 53 were healthy controls.

Demographic and baseline characteristics are shown in Table 1. Mean age was significantly lower in AG group compared with other groups (all p<0.001). Control group had lower mean age than POAG (p<0.001), and OHT (p=0.001) groups. CCT was significantly higher in controls compared with POAG (p<0.001), and AG (p=0.016) groups. No differences were found in the number of glaucomatous medications after excluding healthy subjects.

Table 1 Demographic and Baseline Characteristics of the Study Groups

Primary Outcome

The mean±SD and values of the mixed linear regression for biomechanical corneal parameters, adjusted for age, gender, GAT-IOP, and prostaglandin analogue medication, are shown in Tables 2–4.

Table 2 The Mean Values and Mixed Linear Model (Control Group Set as Reference) for Corvis ST Parameters Adjusted for Age, Gender, GAT-IOP, and Prostaglandin Analogue Medication

Table 3 Mixed Linear Model (Primary Open-Angle Glaucoma Group Set as Reference) for Corvis ST Parameters Adjusted for Age, Gender, GAT-IOP, and Prostaglandin Analogue Medication

Table 4 Mixed Linear Model (Amyloidotic Glaucoma Group Set as Reference) for Corvis ST Parameters Adjusted for Age, Gender, GAT-IOP, and Prostaglandin Analogue Medication

Compared with controls (Table 2), AG eyes had smaller radius (p=0.025), lower HCDeflA (p=0.019), and higher 1/R (p=0.014). There was a trend for a lower SSI (p=0.09) in AG eyes, and a higher SPA1 (p=0.07) in OHT eyes. OHT eyes had also higher SSI (p=0.12), and lower 1/R (p=0.33) than controls, but without reaching statistical significance. No differences were found between POAG eyes and healthy controls.

Compared with POAG eyes (Table 3), OHT eyes had higher SPA1 (p=0.043), and AG eyes had higher 1/R (p=0.010) and DA ratio (p=0.025). There was a trend for a higher HCDeflA in AG eyes (p=0.052) compared with POAG eyes.

Compared with AG eyes (Table 4), OHT had a lower HCDeflA (p=0.028), 1/R (p=0.004), and DA ratio (p=0.01), and a higher SSI (p=0.049).

These results are consistent with more deformable corneas in AG eyes, and less deformable corneas in OHT eyes.

Secondary Outcome

In this analysis, GAT-IOP, GAT-IOPadj, and bIOP were compared. CCT and bIOP were provided by Corvis ST.

As shown in Table 1, GAT-IOP was not significantly different between groups (p=0.171). The same was observed for GAT-IOPadj (p=0.496), and bIOP (p=0.219).

For each group, the differences between GAT-IOP, GAT-IOPadj, and bIOP are demonstrated in Figure 1. Biomechanically corrected IOP was significantly lower than GAT-IOP in POAG group (13.00±4.71 vs 13.36±2.97 mmHg, p=0.005), and control group (13.03±2.36 vs 14.02±2.18 mmHg, p=0.013). The only group in which bIOP was higher than GAT-IOP was AG group (13.68±3.53 vs 12.91±2.90 mmHg), but without reaching statistical significance (p=0.46). In addition, bIOP was significantly lower than GAT-IOPadj in POAG group (13.00±4.71 vs 14.43±3.51 mmHg, p=0.01).

Figure 1 The mean values of intraocular pressure (IOP) obtained by Goldmann applanation tonometry (GAT-IOP), GAT-IOP corrected for pachymetry (GAT-IOPadj), and biomechanically corrected IOP (bIOP) in study groups: control group, primary open-angle glaucoma (POAG) group, amyloidotic glaucoma (AG) group, and ocular hypertension (OHT) group. *p<0.05.

Discussion

Corneal biomechanics has become an interesting topic in glaucoma. The glaucoma disease has a significant prevalence worldwide,¹⁶ and the search for a deeper understanding of glaucoma pathophysiology is ongoing. Our study was conducted to use recently developed technology in the evaluation of corneal biomechanical behavior in patients with POAG, OHT and AG. AG was included in this study because this type of secondary glaucoma is relatively prevalent in endemic regions for ATTR, like Portugal or Sweden, and the underlying mechanisms are very poorly understood.

Our findings were suggestive of more deformable corneas in AG eyes, and less deformable corneas in OHT eyes. Noteworthy, our results were obtained after adjusting for potential influencing factors of corneal biomechanics, like age, gender, GAT-IOP and prostaglandin analogue medication.^17–20 Furthermore, there were significant differences between bIOP and GAT-IOP, which is the gold standard method for evaluating IOP, or GAT-IOPadj.

Many studies on corneal biomechanics in glaucoma have suggested that POAG corneas were less deformable than normal corneas.^21–23 A lower corneal hysteresis (CH) obtained with ORA device was found in POAG eyes which reflects a lower ability of the cornea to absorb and dissipate the mechanical energy applied by a deformation force.²¹ Using the Corvis ST, POAG eyes had lower deformation amplitude compared with healthy subjects.^22–24 Wang et al,²² and Lee et al²⁵ found a greater A2V and PD (indicating a stiffer cornea) in glaucomatous eyes. In contrast, Miki et al²⁶ reported a smaller A1T and Rad consistent with more deformable corneas in medically controlled glaucomatous eyes and the lack of control for confounding factors in previous studies was pointed out as the cause for the discrepant results. In agreement with our study, Pradhan et al²⁷ demonstrated a similar corneal biomechanical profile between POAG patients and healthy subjects. Eyes under anti-glaucoma medication were excluded from this study. In fact, there is evidence that prostaglandin analogues affect the corneal biomechanical properties by activation of matrix metalloproteinases, resulting in a lower CH and higher deformation amplitude.^19,20 In our work, as in that by Pradhan et al,²⁷ the effect of prostaglandin analogues in corneal biomechanics was taken into consideration, which may explain the similar results. Additionally, the present study is one of the few that analyzed the novel mathematically-derived parameters (DA ratio, 1/R, SPA1, and SSI) in corneas of glaucoma eyes.^12,28 It was proposed that these parameters were less dependent of IOP,^29,30 and correlated with corneal biomechanics (including CH).²⁸ When these novel parameters were assessed, POAG group did not display a different corneal biomechanical behavior compared with controls, which is in line with previous published research.^12,28

OHT has also been studied with this recent technology. A higher CH was documented in patients with OHT as compared to those with glaucoma.^31,32 A recent study demonstrated a higher A1DeflL and A2V, and lower A1V compatible with stiffer corneas in OHT eyes, while we found a higher SSI.³³ The differences in corneal biomechanics between OHT eyes and those with glaucoma, either POAG or AG, were evident in the present study. Moreover, OHT eyes usually present a thicker CCT.³² Thus, eyes with OHT can be less susceptible to ONH damage due to biomechanical properties of ocular structures, including peripapillary sclera. A thicker and stiffer peripapillary sclera can be a protective feature of ONH damage in OHT eyes.

AG is a secondary open-angle glaucoma that develops in patients with ATTR. Although glaucoma affects about 20% of ATTR patients,^34,35 this disease has been little explored in the research. For unknown reasons, the course of AG is usually accelerated and often requires surgical treatment.^36,37 Importantly, we found thinner and more deformable corneas in AG eyes which may represent an increased susceptibility to ONH damage. Hence, the rapid progression of AG may be related to the presence of a thinner and more deformable peripapillary sclera. Recent data suggested that more compliant corneas have a greater risk of glaucoma progression.³⁸ In addition, bIOP was on average higher than GAT-IOP only in AG eyes, which may have important clinical implications. We hypothesize that, in this subset of patients, it may be reasonable to measure bIOP when there is rapid progression of glaucoma despite apparently controlled IOP. It is possible that biomechanical properties of amyloidotic eyes are at least partially responsible for glaucoma onset and aggressive progression.

To summarize, this study did not corroborate previous research which demonstrated less deformable corneas in POAG eyes compared with controls. However, OHT corneas seem to be less deformable than POAG corneas, as shown by a higher SPA1. In contrast, AG corneas seem to be more deformable than POAG corneas, as demonstrated by a higher 1/R and DA ratio.

In agreement with a study by Vinciguerra et al,¹² bIOP was significantly lower than GAT-IOP and GAT-IOPadj in POAG patients. Likewise, GAT-IOP adjusted for CCT and bIOP do not seem to be interchangeable. Corneal compensated IOP (IOPcc) is a biomechanical parameter assessed by the ORA device, comparable to bIOP provided by the Corvis ST. Two studies^39,40 reported a higher value of IOPcc compared with GAT-IOP. This led us to note the disparity between the comparison of GAT-IOP with these two types of biomechanically corrected IOP, IOPcc (ORA device) and bIOP (Corvis ST device). Matsuura et al⁴¹ corroborated these findings showing that bIOP is lower than IOPcc and GAT-IOP. In intracameral studies,^42–44 true IOP was usually higher than GAT-IOP, such as IOPcc, but not bIOP. Thus, there may not be a relationship between bIOP and true IOP. However, an ex vivo study showed no differences between bIOP and internal manometric IOP (0.3±1.6 mmHg, p=0.989) in 5 ocular globes of humans.⁴⁵

This study has limitations. First, the cross-sectional nature of our study design prevented the evaluation of causality or prospective prediction of glaucoma risk. All glaucoma patients were under topical treatment, and, to minimize the effect of prostaglandin analogues in corneal biomechanics, it was regarded as a confounding factor in mixed linear regression. Of note, we performed a statistical correction for the inclusion of both eyes in some study participants. Some subjects were previously submitted to ophthalmic surgery, but a minimal interval of 6 months was imposed by inclusion criteria to overcome this limitation. The influence on the results is expected to be minimal after this period.^46,47 Afterwards, it was not possible to assess the glaucoma severity due to unavailability of automated perimetry at the time of the study. Finally, it should be noted that the sclera and cornea differ on the dimensions and arrangement of collagen fibrils, which may lead to distinct biomechanical behavior.⁴ This underlines the importance of in vivo methods for assessing scleral and LC structural properties to become available.

As strengths of this study, we underline the interesting and novel findings in the group of AG eyes. To our knowledge, no other study has analyzed simultaneously so many corneal biomechanical parameters. This comparative study exposed the differences in biomechanical behavior of eyes with POAG, OHT and AG, and supports the role of ocular biomechanics in individual susceptibility to glaucomatous damage. Therapeutic approaches aiming to modify biomechanical properties of peripapillary sclera have emerged, and this kind of studies will help to establish a rationale for them.⁴⁸ Moreover, different IOP measurements provided different values, and it became evident that further investigation is needed to support the use of IOP corrected for corneal biomechanics.

To conclude, our results were suggestive of more deformable corneas in AG eyes, and less deformable corneas in OHT. Thus, AG eyes seem to be more susceptible to glaucomatous damage and progression, while OHT eyes seem to be more resistant to ONH damage. There were significant differences between GAT-IOP, GAT-IOP adjusted for CCT and bIOP. Further investigation is needed to validate the use of bIOP.