One hundred and forty-six eyes of 90 glaucoma patients were recruited with follow-up duration (mean: 6.76 ± 0.65 years; range: 4.56-7.61 years). All patients were followed at approximately 4-month intervals during the period from June 2007 to December 2014 at the University Eye Centre, the Chinese University of Hong Kong. The detailed protocols in accordance with the Declaration of Helsinki were approved by the institutional review board of The Chinese University of HongKong. All subjects were told of the purpose of the study and gave written informed consent before inclusion. At each follow-up visit, subjects underwent a full ophthalmic examination, visual field examination, RNFL imaging with OCT, and optic nerve head with CSLO. We confirm that all methods were carried out in accordance with clinical research guidelines and regulations of The Chinese University of HongKong. All included eyes had best corrected visual acuity of at least 20/40. Eyes with retinal pathology, macular disease and refractive or retinal surgery were excluded. Eyes with uncomplicated cataract or glaucoma filtration surgery could be included in this study. Patients with glaucoma were identified on the basis of the presence of visual field defects with corresponding optic disc and RNFL changes in at least 1 eye independent of the level of IOP. During the follow-up, patients were treated at the discretion of the attending ophthalmologists with reference to the target IOP. Progressive RNFL thinning and ONH surface depression was detected with GPA (Carl Zeiss Meditec) and TCA (Heidelberg Engineering), respectively.
Corneal Hysteresis Measurement with Ocular Response Analyzer
Corneal hysteresis was acquired with ocular response analyzer (ORA, Reichert Inc, Depew, NY). For ORA measurement, a metered air pulse is delivered to the cornea and the cornea is flattened at two different time points (force-in applanation and force-out applanation). An electro-optical collimation detector system is used to record the light reflectivity from the cornea. The reflectivity is most intense when the cornea is completely flattened. CH is derived from the difference between the force-in and force-out applanation . Three other parameters were also measured from each ORA signal profile: IOPcc, IOPg and CRF. Four ORA measurements were taken on an examined eye each time and the average value was calculated; only measurements with a waveform score larger than 5 were accepted.
IOP Measurement with Dynamic Contour Tonometry
IOP measurement was taken by dynamic contour tonometry (DCT) (Pascal; Swiss Microtechnology AG, Port, Switzerland). The working principle of DCT has been described . In brief, the matching of the contour of the tonometer and the contour of the cornea allows the IOP measured by a pressure sensor located at the tip of the tonometer. In this study, the recording duration of each measurement was about 5 seconds. Measurement with a quality score less than 3 (scale 1 to 5) was repeated until 2 measurements of quality score between 1 and 3 (as recommended by the manufacturer) were obtained in each eye and the average value was calculated. DCT measures the systolic IOP and diastolic IOP and only diastolic IOP was analyzed in the study.
Optical Coherence Tomography Imaging
Spectral domain OCT imaging was performed with Cirrus HD-OCT (software version 5.0; Carl Zeiss Meditec). The acquisition rate of the Cirrus HD-OCT was 27,000 A-scans per second and the transverse and axial resolutions were 15μm and 5μm, respectively. An optic disc cube consisting 200×200 axial scans, 6×6mm2 centered on the optic disc was used to evaluate the optic disc parameters for every subject. Images were captured by operators with at least 1 years’ experience using the Cirrus HD-OCT (Carl Zeiss Meditec). The pupils were not routinely dilated during RNFL imaging. However, dilation with tropicamide 0.5% and phenylephrine 0.5% each was performed when the pupil size was too small for images of the required quality to be obtained. Images with poor centration, motion artifact, poor focus, or missing data were detected by the operator at the time of imaging, with rescanning performed in the same visit. Each OCT scan included in the study had signal strength ≥7. Saccadic eye movement was detected with the line-scanning ophthalmoscope overlaid with OCT en face during OCT imaging. Images with motion artifact were rescanned in the same visit [22,23]. Serial RNFL thickness maps were analyzed for detection of change using the Guided Progression Analysis (GPA) .
The Cirrus HD-OCT Guided Progression Analysis (GPA) (Carl Zeiss Meditec AG) was used to analyze serial RNFL thickness maps (200×200 pixels) for detection of progressive RNFL thinning. Guided Progression Analysis automatically aligned and registered 2 baseline and the follow-up OCT images so that the same superpixel (1 superpixel=4×4 pixels) locations could be analyzed for detection of change. The difference in RNFL measurement of an individual superpixel between the baseline and the follow-up RNFL thickness maps was compared with an estimate of test-retest variability of that particular superpixel (proprietary database from Carl Zeiss Meditec AG). Superpixels with an RNFL measurement difference exceeding the test-retest variability between a follow-up and the first and second baseline images would be encoded in yellow in the OCT RNFL thickness change map (50×50 superpixels). If the same changes were evident in an additional consecutive follow-up image, the superpixels would be encoded in red. In this study, the 2 baseline images were separated by approximately 4 months and progressive RNFL thinning was confirmed when an area of more than 20 superpixels (factory default) was encoded in red in the RNFL thickness change map for at least 2 consecutive visits. At least 3 consecutive follow-up visits showing significant RNFL thickness reduction were required to confirm progressive RNFL thinning.
Confocal Scanning Laser Ophthalmoscopy Imaging
Optic disc imaging was performed with the HRT 3 (software version 3.0, Heidelberg Engineering). A three dimensional topographic image consisting of up to 384 x 384 x 64 pixels was constructed from multiple focal planes axially along the optic nerve head. An average of three consecutive scans was obtained and aligned to compose a single mean topography for analysis. The optic disc margin was outlined by an experienced examiner on the mean topographic image. Once the contour line was drawn, the software automatically calculated all the optic disc measurements. The reference plane is defined at 50μm posterior to the mean retinal height between 350° and 356° along the contour line. The area above and below the reference plane is defined as rim and cup respectively. Re-scanning was performed in the same visit if motion artifacts were detected immediately after the imaging. All eyes included in the analysis had an image quality standard deviation ≤30µm.
The HRT Topographic Change Analysis (TCA, Heidelberg Engineering) was used to analyze serial ONH topography images (96×96 superpixels; 1 superpixel =4×4 pixels) for detection of ONH surface depression . Individual superpixel ONH surface height measurements were compared between the baseline and each of the follow-up examinations with an F test. The pooled variability of the baseline and the follow-up examinations of a particular pixel was compared with the within variability of the baseline and the follow-up examinations (with an error probability of the F-test <5%). If significant ONH surface depression was detected in a superpixel and confirmed with at least 2 consecutive follow-up visits, the superpixel would be encoded in red in the significance map. The saturation of the color increased with the magnitude of surface height change. Progressive ONH surface depression was defined using 3 criteria (liberal, moderate, and conservative) with reference to the area and depth of ONH surface depression adopted from the studies by Chauhan and colleagues [26,27]. The liberal criterion required a cluster of ≥0.5% of the disc area and a depth change of ≥20 mm; the moderate criterion a cluster of ≥1% of the disc area and a depth change of ≥50 mm; and the conservative criterion a cluster of ≥2% of the disc area and a depth change of ≥100 mm. In this study, the moderate criterion was adopted to ensure a fair specificity for ONH surface depression detection with HRT and RNFL thinning acquired with OCT . The ONH surface depression was defined when the change was detected at least 3 of 4 consecutive follow-up examinations.
Visual Field Examination
Visual field was obtained with the white-on-white SITA standard 24-2 strategy in Humphrey Field Analyzer II-i (Carl Zeiss Meditec). All visual fields included in the study had fixation losses, false positive and false negative errors less than 20%. Average visual field sensitivity was expressed in MD (mean deviation), as calculated by the perimetry software. A visual field defect was defined as having three or more significant (P<0.05) non-edge contiguous points with at least one at the P<0.01 level on the same side of horizontal meridian in the pattern deviation plot and confirmed with at least two consecutive examinations [22,23].
Visual field progression was analyzed with event-based analysis using the Guided Progression Analysis (GPA, Carl Zeiss Meditec) according to EMGT criteria . Progression was defined when there were ≥3 points that showed significant changes (greater than the test-retest variabilities) compared with two baseline examinations (separated by ~4 months in this study) for ≥2 or 3 consecutive tests (i.e. “likely progression” was noted in the GPA printout in the latest follow-up visit).
Statistical analyses were performed using Stata version 12.0 (Stata Corp LP, College Station, TX). Differences in demographics, visual field, ONH, and RNFL measurements between the progressive and nonprogressive eyes were compared with linear mixed models after adjustment of correlation between fellow eyes. A multivariate cox proportional hazards model was used to investigate if baseline corneal hysteresis and other risk factors including baseline age, CCT, baseline MD, baseline disc area, baseline diastolic IOP, and mean IOP during follow-up were risk factors for progressive RNFL thinning, ONH surface depression and visual field progression. Adjusted HRs from the multivariate cox proportional hazards models were reported in this study. P < 0.05 was considered statistically significant.