To analyze and compare corneal biomechanical properties in healthy black and white subjects using the Ocular Response Analyzer (ORA) and to evaluate their relationship with other ocular parameters.
Observational cross-sectional study.
One hundred eighty one eyes (46 in blacks, 135 in whites) of 119 patients (37 blacks, 82 whites) were recruited from the longitudinal Diagnostic Innovations in Glaucoma Study (DIGS) and from the African Descent and Glaucoma Evaluation Study (ADAGES) at the University of California, San Diego. Corneal curvature, axial length, central corneal thickness (CCT), corneal hysteresis (CH), and corneal resistance factor (CRF) were obtained from all participants. Univariable and multivariable regression analyses were used to evaluate the associations between ORA measurements and age, CCT, axial length, corneal curvature, and race.
Black subjects had significantly lower values of CH (9.7 mm Hg vs 10.4 mm Hg; P = .033), CRF (9.84 mm Hg vs 10.70 mm Hg; P = .028), and CCT (534 μm vs 562 μm; P = .001) compared to white subjects. A significant relationship was found between CH and CCT (R 2 = 0.25; P < .001) and between CRF and CCT (R 2 = 0.42; P < .001). After adjusting for CCT, age, axial length, and corneal curvature, the difference between blacks and whites in CH ( P = .077) and CRF ( P = .621) measurements lost statistical significance.
Black subjects tended to have lower measurements of corneal hysteresis compared to white subjects; however, this was largely explained by differences in corneal thickness. Therefore, it is unlikely that CH would have an independent effect in explaining differences in susceptibility of disease between these 2 racial groups.
Goldmann Applanation Tonometry (GAT) has been considered the reference standard for intraocular pressure (IOP) measurement for several decades. However, there are numerous sources of error that may significantly influence IOP readings obtained with GAT, notably corneal thickness. IOP may be overestimated or underestimated in eyes with thick or thin corneas, respectively, resulting in misclassification and mismanagement of a significant number of glaucoma patients and subjects with ocular hypertension. In addition, several clinical trials have shown that corneal thickness is a risk factor for development as well as progression of glaucoma, and there is speculation that this effect may be independent of its role in affecting GAT measurements.
Important differences have been described in corneal thickness measurements in blacks compared to whites. According to clinic-based and population-based studies, blacks have thinner corneas than whites, which can potentially be related to a higher risk of glaucoma development and possibly to the more aggressive rates of disease progression frequently found in the former group. The thickness of the cornea, however, is just 1 among several corneal physical properties that influence the measurement of IOP with applanation tonometry. Other biomechanical parameters such as elasticity or viscoelastic properties may also influence corneal resistance to applanation and, therefore, IOP measurements obtained by GAT. Previous investigators have shown that corneal hysteresis (CH), a corneal biomechanical property related to viscoelastic dampening, is reduced in glaucomatous compared to healthy eyes and is a risk factor for glaucoma progression that appears to be independent of corneal thickness. In addition, a recent study suggested that CH, but not corneal thickness, is associated with optic disc surface compliance, which may ultimately be related to glaucoma pathogenesis. Therefore, it is possible that differences in CH could potentially be related to differences in disease outcomes in blacks compared to whites.
Despite the potential relevance of CH, no study has yet compared CH measurements in blacks and whites. The Ocular Response Analyzer (ORA; Reichert Inc, Depew, New York, USA) measures corneal biomechanical properties such as CH and corneal elasticity by analyzing corneal responses submitted to air jet–induced deformation. The purpose of this study was to analyze and compare corneal biomechanical properties using the ORA in a healthy population of blacks and whites, and to investigate their relationship with other ocular parameters.
This was an observational cross-sectional study. All participants included in this study were retrospectively selected from the longitudinal Diagnostic Innovations in Glaucoma Study (DIGS) and the African Descent and Glaucoma Evaluation Study (ADAGES) conducted at the Hamilton Glaucoma Center, University of California, San Diego. Briefly, these studies were designed to evaluate optic nerve structure, visual function, and risk factors in glaucoma. For this particular study, we have included a cohort of healthy participants recruited by advertisement from the general population. Participants were asked to identify their racial category by self-reporting using the National Eye Institute inclusion/enrollment system ( orwh.od.nih.gov/pubs/outreach.pdf [page 22]).
All individuals underwent a complete ophthalmologic examination, including visual acuity assessment, slit-lamp biomicroscopy, gonioscopy, dilated fundoscopic examination using 78-diopter (D) lens, stereoscopic disc photography, and automated perimetry using the 24-2 Swedish Interactive Threshold Algorithm (SITA; Carl Zeiss Meditec, Inc, Dublin, California, USA). Central corneal thickness (CCT) was measured using an ultrasound pachymeter (Pachette DGH 500; DGH Technology, Inc, Philadelphia, Pennsylvania, USA) over an undilated pupil and the mean of 3 readings was recorded. Corneal curvature was obtained using an autorefractor (Humphrey – Zeiss model S97; Carl-Zeiss Meditec). Axial length was acquired with IOLMaster (Carl-Zeiss Meditec).
All eyes included in this study had best-corrected visual acuity of 20/40 or better, normal fundus examination results with a healthy appearance of the optic disc and retinal nerve fiber layer, open angle at gonioscopy, and normal visual field. A normal visual field was defined as a mean deviation and a pattern standard deviation within the 95% normal confidence limits and glaucoma hemifield test results within normal limits. Participants were excluded if they had previous use of anti-glaucoma medication or any previous ocular surgery. Intraocular pressure was not used as an inclusion or exclusion criterion.
Corneal biomechanical properties were measured using software version 2.02 of the ORA. Briefly, the ORA uses an air pulse to flatten the corneal surface, causing the cornea to shift inward, passing from a flat to a concave state. As the air pulse decreases, the cornea returns first to a flat state and then to its initial convex shape. An electro-optical collimator-detector records the 2 applanation events (P1 and P2) produced by the bidirectional movement of the cornea. The time difference between the first and the second applanation is approximately 20 milliseconds, short enough to ensure that eye position or ocular pulse does not change during examination.
The device provides 4 different readings. CH is defined as the numerical difference between the 2 applanation states ( P1 minus P2 ), being an indicator of viscoelastic dampening. Corneal-compensated IOP (IOPcc) is an IOP measurement that minimizes corneal influence. It is calculated as (P1-k) times P2 , where k is a constant derived from previous studies evaluating eyes that underwent corneal change induced by laser in situ keratomileusis (LASIK). IOP Goldmann (IOPg) is the Goldmann applanation tonometer correspondent in this device, and is calculated as the average of P1 and P2. Corneal resistance factor (CRF) is a measurement of another corneal biomechanical property, the resistance to deformation. To ensure quality of the readings, images were reviewed and images with applanation peaks on the ORA waveform that were fairly symmetrical in height were included. Also, we included only measurements with a wavelength score of 7 or higher marked as best signal value. We obtained at least 2 measurements per eye and selected the one with the highest quality score for the analysis. These scores are a direct reflection of signal quality and reliability of measurements. Although strict cut-off points for wavelength score have not yet been determined in the literature, a score over 7 has been suggested by the manufacturer (David Taylor, Reichert Inc, personal communication) as indicative of good quality.
Initially, we used univariable models to evaluate differences in ocular biomechanical properties and other ocular parameters between blacks and whites. Subsequently, we built multivariable models to evaluate differences in corneal biomechanical parameters, ie, CH and CRF, in blacks versus whites while adjusting for differences in other parameters such as age, corneal thickness, axial length, and corneal curvature. The multivariable models were built in 2 steps. First, differences in CH and CRF between blacks and whites were adjusted for age, corneal thickness, axial length, and corneal curvature; simultaneously, in a full model. Then, a reduced multivariable model was built adjusting only for covariates that significantly influenced CH and CRF, based on results from the full model.
In addition, we studied the association among corneal properties (CH, CRF, CCT, axial length, and curvature) using linear regression models.
Generalized estimating equations with robust standard error in a Huber-White matrix sandwich were used to adjust for potential correlations between both eyes of the same individual.
Statistical analyses were performed with commercially available software (STATA v. 10.0, StataCorp, College Station, Texas, USA; and SPSS v.16.0, SPSS Inc, Chicago, Illinois, USA). A P value less than .05 was considered statistically significant.
The study included a total of 181 eyes (135 in whites, 46 in blacks) of 119 participants (82 whites, 37 blacks). The mean (standard deviation) age of the included subjects was 64.13 (13.33) years, ranging from 24 to 90 years. On average, white subjects were older than blacks (66 years vs 58 years; P = .009).
Table 1 shows the results of the comparison of ocular parameters between blacks and whites. Mean CCT was significantly lower in blacks compared to whites (534 μm vs 562 μm; P = .001). Blacks had significantly lower CH values compared to whites (9.7 mm Hg vs 10.4 mm Hg; P = .033) and CRF measurements were also significantly lower in blacks than in whites (9.84 mm Hg vs 10.70 mm Hg; P = .028). No statistically significant difference between blacks and whites was found for axial length, corneal curvature, IOPg, IOPcc, and GAT measurements.
|Parameter||Whites (n = 135 eyes)||Blacks (n = 46 eyes)||P Value|
|Age (years)||66.00 ± 11.9||58.41 ± 15.7||.009|
|CCT (μm)||562 ± 36.8||534.09 ± 39.4||.001|
|GAT (mm Hg)||16 ± 2.9||15.77 ± 2.81||.658|
|AL (mm)||23.90 ± 1.21||23.77 ± 1.15||.579|
|Corneal curvature (D)||43.28 ± 1.68||43.76 ± 1.65||.148|
|CH (mm Hg)||10.44 ± 1.6||9.70 ± 1.72||.033|
|CRF (mm Hg)||10.70 ± 1.8||9.84 ± 1.92||.028|
|IOPg (mm Hg)||16.36 ± 4.35||15.57 ± 3.37||.273|
|IOPcc (mm Hg)||16.69 ± 4.2||16.80 ± 3.16||.874|
No statistically significant association was found between CCT and both axial length ( P = .569) and corneal curvature ( P = .496). Axial length was significantly associated with corneal curvature ( P < .001). For each 1-mm increase in axial length, the curvature was reduced by 0.82 D.
Figure 1 shows a locally weighted scatterplot smoothing (LOWESS) of CH versus CCT measurements according to racial group. The observed relationship was close to linear. A significant association was seen between CH and CCT, with lower CH measurements in eyes with thinner corneas, for both black (R 2 = 0.41; P < .001) and white subjects (R 2 = 0.16; P < .001), with no statistically significant difference between the 2 groups ( P = .08, for interaction variable). Considering all participants from both racial groups, each 100-μm increase in corneal thickness was associated with a 2.14-mm Hg increase in CH.
A multivariable model was then built to evaluate differences in CH measurements between the 2 racial groups ( Table 2 ). Only age and CCT were significantly associated with CH. Differences in CH measurements between blacks and whites were not statistically significant in the full multivariable model ( P = .07) as well as in a reduced multivariable model adjusting only for age and CCT ( P = .120).