Determinants of posterior corneal biometric measurements in a multi-ethnic Asian population.

Marcus Ang,Tin Aung,Tien Yin Wong,Wesley Chong,Jodhbir S Mehta,Ming-Guang He,Huiqi Huang,Rajiv R Mohan

doi:10.1371/journal.pone.0101483

Abstract

PurposeTo describe the corneal and anterior segment determinants of posterior corneal arc length (PCAL) and posterior corneal curvature (PCC).MethodsCross-sectional, population-based study of 1069 subjects (1069 eyes) aged 40–80 years, from three major Asian ethnic groups. All underwent anterior segment optical coherence tomography imaging and analysis with Zhongshan Angle Assessment Program. Our main outcome measures were determinants of PCAL and PCC using adjusted, multivariate linear regression analysis, adjusted for confounders to obtain the estimated marginal means (EMM) with standard error (SE).ResultsThe overall mean (± SD) of PCC was: 6.51±0.39 mm; and PCAL was: 12.52±0.59 mm. Malays had a relatively longer PCAL (EMM = 12.74 mm, SE = 0.04 mm) than Chinese (EMM = 12.48 mm, SE = 0.03 mm, P<0.001), and Indians (EMM = 12.42 mm, SE = 0.03 mm, P<0.001). Anterior segment parameters had weak-moderate correlations with PCAL, which included: anterior chamber depth (ACD) (r = 0.55, P<0.001), PCC (r = 0.27, P<0.001), anterior corneal curvature (ACC) (r = 0.14, P<0.001) and central corneal thickness (CCT) (r = −0.07, P = 0.023). In multivariate analysis, anterior segment parameters explained only 37.6% of the variance of PCAL, with ACD being the most important determinant (partial R2 = 0.300; P<0.001). The determinants of PCC included ACC, PCAL and CCT (explaining 72.1% variation of PCC), with ACC being the most important determinant (partial R2 = 0.683; P<0.001).ConclusionThere was moderate correlation of PCAL with ACD, but anterior segment parameters accounted for only a small proportion of the variation in PCAL. The significant differences in PCAL and PCC amongst different Asian ethnic groups suggests that there is a need to consider this factor when planning for anterior segment surgeries such as endothelial keratoplasty.

Highlights

Accurate measurement of the posterior corneal surface is important in the planning and execution of endothelial keratoplasty, which is increasingly becoming the surgery of choice in endothelial disease.[1]
These differences may have impact on our clinical practice, such as intra-operative graft sizing for endothelial keratoplasty, where currently measurements are based on visual estimation of the horizontal white-to-white diameter that does not take into account the cornea curvature.[20]
We found that anterior corneal curvatures (ACC), posterior corneal arc length (PCAL), and central corneal thickness (CCT) were the most significant determinants of posterior corneal curvature (PCC)

Summary

Introduction

Accurate measurement of the posterior corneal surface is important in the planning and execution of endothelial keratoplasty, which is increasingly becoming the surgery of choice in endothelial disease.[1] One way to more precisely image the cornea is with the use of anterior segment optical coherence tomography (AS-OCT, Visante; Carl Zeiss Meditec, Dublin, CA).[2,3] measurements of anterior segment parameters from AS-OCT images are often inaccurate due to subjective placement of calipers or measurement tools.[4] The Zhongshan Angle Assessment Program (ZAAP, Guangzhou, China) has been shown to reliably derive these parameters from AS-OCT images,[5,6,7,8] with high inter- and intra-observer agreement.[5,6] The ZAAP software works by extracting the gray-scale images from the AS-OCT images and uses image processing and algorithms to define the anatomical landmarks.[5] This enables users to obtain rapid and objective analysis of anterior segment images for surgical planning and diagnosis of various ocular diseases (ZAAP research software available upon request; none of the authors have any commercial interest).[2,3,9] The ZAAP analysis obtains standard parameters such as central corneal thickness (CCT), anterior chamber depth (ACD), anterior corneal curvatures (ACC) using image-processing algorithms. [5]

Methods

Results

Conclusion