Aim To evaluate quantitative changes in anterior segment optical coherence tomography (AS-OCT) parameters at 4 years in subjects with open angles on gonioscopy at baseline.
Methods 339 gonioscopically open-angle subjects aged >50 years underwent AS-OCT imaging (Visante, Carl Zeiss Meditec, Dublin, California, USA) at baseline and at 4 years. Customised software was used to analyse images. Linear regression was performed to assess baseline predictors of change in mean trabecular iris space area (TISA).
Results Of the 339 subjects, 204 (61%) had good images for analysis. The mean age at baseline was 60 (SD 6.42) years; 55.4% were women and 87.7% were Chinese. Overall, there was a decrease in anterior chamber depth (ACD), anterior chamber area (ACA), anterior chamber volume (ACV) and angle opening distance at 500/750 µm, TISA 500/750 µm, iris thickness at 2000 µm and iris area in both nasal and temporal quadrants (p<0.05) at 4 years. There was an increase in iris curvature (ICURV) and lens vault (LV) (p<0.05). With univariate analysis, shorter axial length (AxL), shallower ACD, greater LV, smaller ACA/ACV and larger mean ICURV at baseline were associated with less change in TISA750 at follow-up. On multivariate analysis, only baseline mean ICURV and shorter AxL were predictive of less change in TISA750 at 4 years.
Conclusions Angle width significantly decreased; ICURV and LV increased in subjects with gonioscopically open angles followed at 4 years. Eyes with shorter AxL and greater ICURV at baseline had lesser decrease in angle width.
Statistics from Altmetric.com
Primary angle-closure glaucoma is a major cause of irreversible blindness in Asia, especially in people of Chinese ethnicity.1 ,2 Anterior segment optical coherence tomography (AS-OCT) provides anterior segment cross-sectional images in real time allowing for quantitative measurements. Using AS-OCT, we have identified novel parameters associated with angle closure, such as smaller anterior chamber area (ACA) and anterior chamber volume (ACV) and greater lens vault (LV),3–6 as well as iris parameters such as iris curvature (ICURV), iris cross-sectional area (IAREA) and iris thickness (IT).4 ,7–9 However, little is known about changes in these anterior segment parameters that occur over time with ageing.
In 2007, we conducted a community-based study of 2114 subjects that evaluated the use of AS-OCT in screening for narrow angles in a Singapore population. When compared with gonioscopy, AS-OCT was found to have a sensitivity of 88.4% and specificity of 62.9% in diagnosing angle closure.10
In the current study, we reimaged participants in the screening study with open angles at baseline (as assessed by gonioscopy) using AS-OCT, 4 years after the initial exam. We present the anterior segment changes over time in this population.
Study population selection
The primary study population consisted of phakic subjects aged 50 years or older who were examined in a community-based study of Singaporeans visiting a government-based polyclinic for non-ophthalmic medical problems in 2007. The study methodology and details have been described previously.10
In the current study, 585 subjects with open angles on gonioscopy (ie, open on all quadrants with visible posterior trabecular meshwork) at baseline were recalled 4 years after the screening evaluation for clinical and AS-OCT assessments. Of these subjects, 485 had one or more quadrants closed on AS-OCT (iris–trabecular contact beyond scleral spur in any of the four quadrants) at baseline and the remaining 100 had open angles on both gonioscopy and AS-OCT. The latter group was chosen by a computer-generated random selection from the overall 738 subjects (with open angles on all quadrants with visible posterior trabecular meshwork) eligible from the baseline cohort. After an initial medical and ophthalmic history interview, all respondents underwent the following examinations: visual acuity, anterior segment imaging by AS-OCT in the dark (Visante AS-OCT, Carl Zeiss Meditec, Dublin, California, USA) and anterior chamber depth (ACD) and axial length (AxL) measurements (IOLMaster—software V.3.02, Carl Zeiss Meditec, Dublin, California). AS-OCT and IOLMaster versions were the same at baseline and at 4-year follow-up study. Individuals at baseline and follow-up were excluded if they had a history of intraocular surgery, previous anterior segment laser treatment, penetrating ocular trauma, aphakia/pseudophakia, corneal disorders that could influence AS-OCT imaging such as corneal endothelial dystrophy, corneal opacity or severe pterygium during the past four years.
Imaging with AS-OCT was performed in dark room conditions (0 lux) by a single operator who was masked to the results of other tests. The standard anterior segment single-scan protocol, which produces 256 scans in 0.125 s, was used and scans were centred at the pupil. To obtain the best-quality image, the examiner adjusted the saturation, noise and optimised the polarisation for each scan. The examiner chose the best image with the least motion or image artefacts resulting from the eyelids. Similar lighting conditions, scan protocols and analysis were used for baseline and follow-up.
The Zhongshan Angle Assessment Program (ZAAP; Zhongshan Ophthalmic Center, Guangzhou, China) was used to analyse the AS-OCT images for quantitative measurements of the anterior segment (figure 1).11 The same software version was used for both baseline and follow-up image analysis. A single observer (YH) who was masked to the clinical data analysed the de-identified baseline and follow-up images. For each image, the only observer input was to determine the location of the 2 scleral spurs (SS). The algorithm then automatically calculated the anterior segment parameters. Anterior chamber width was defined as the horizontal SS-to-SS distance, and LV was the perpendicular distance from the horizontal line between the 2 SS to the anterior pole of the lens.5 Angle opening distance (AOD) at 750 µm from SS (AOD750) was the perpendicular distance from the iris to the trabecular meshwork at 750 µm anterior to the SS.12 Trabecular iris space area at 750 µm from SS (TISA750) was defined as the trapezoidal area with these boundaries: anteriorly, AOD750; posteriorly, a line perpendicular to the plane of the inner corneoscleral wall drawn from SS to the opposing iris; superiorly, the inner corneoscleral wall; and inferiorly, the anterior iris surface. IAREA lies within the iris contour bordered by a line through SS and perpendicular to the meshwork line.13 IT was defined as the shortest distance of the iris at 2000 µm from SS between the anterior and posterior iris surface. ICURV was determined by creating a line from the most peripheral to the pupillary edge of the iris and then measuring the perpendicular distance from this line to the greatest convexity point along the posterior iris surface.14 ACA was defined as the cross-sectional area of anterior segment bounded by endothelium, anterior surface of iris and anterior surface of lens (within the pupil). A vertical axis through the midpoint (centre) of ACA was plotted by the program and ACV was calculated by rotating ACA 360° around this vertical axis.15 The repeatability of anterior segment measurements obtained using ZAAP software has been shown to be excellent, with an intraclass correlation coefficient exceeding 0.88.3 Horizontal scans imaging both nasal and temporal angles only were included in analysis as ZAAP programme can only process images with both angles.12
Measurements of the other ocular variables
A trained ophthalmic technician measured ACD and AxL with the IOLMaster. The IOLMaster measures ACD from the corneal epithelium to the anterior lens surface with lateral slit illumination. The averages of five readings obtained for ACD and AxL, and three consecutive readings obtained for the corneal curvature, were calculated and used for subsequent analyses. All readings for each parameter were required to be within 0.05 mm of the reading within the highest signal:noise ratio. Goldmann applanation tonometry was used to measure intraocular pressure (IOP), and the vertical cup:disc ratio was determined clinically using a 78-D lens at the slit lamp with a graticule (Haag-Streit Model BQ-900; Haag Streit, Koeniz, Switzerland). Median of three IOP readings was used for analysis. All clinical measurements were performed by a glaucoma fellowship trained ophthalmologist (MB) masked to the baseline and follow-up AS-OCT findings.
Statistical analyses were performed using SPSS statistics V.18.0 for Windows (PASW Statistics 18, SPSS, Chicago, Illinois, USA). Continuous variables were described as the mean, SD and range. Paired t test was performed to assess the significance of change between baseline and follow-up measurements (using a p value of 0.05). Demographic and clinical parameters at the baseline evaluation between the respondents and non-respondents were also compared with t test for continuous variables and χ2 test for categorical variables.
Univariate linear regression models were constructed to assess the association of risk factors with mean TISA750 change (baseline average of nasal and temporal TISA750 minus follow-up average of nasal and temporal TISA750). Multiple linear regression models were then constructed to assess the independent effects of the various factors that were significantly associated with mean TISA750 change in age-adjusted and gender-adjusted univariate analysis. In the multivariate model, the means of the nasal and temporal parameters were used for ICURV and IT2000 in addition to AxL, ACD and LV. Unstandardised regression coefficient (β), its 95% CI and the significance level were obtained for individual variables.
Telephone calls were made to a total of 585 subjects, 56 were ineligible (23 of whom had bilateral cataract surgery over the follow-up period and 33 had morbid medical conditions such as stroke and heart attack) and 12 had died. Of the remaining 517 subjects, 339 (65.6%) responded and were able to make a follow-up visit, 118 (22.8%) refused to come in for a follow-up visit and 60 (11.6%) were uncontactable (figure 2).
Men were less likely to participate (59.7% vs 40.3%, p=0.003), but there were no significant differences in age, ethnicity and other ocular clinical parameters such as AxL, ACD, IOP and mean Shaffer grading between the respondents and non-respondents (table 1). While 339 subjects participated, complete data analysis was possible in 204 subjects (61%) due to inability to clearly identify SS in the horizontal AS-OCT images taken at both baseline and follow-up in the remaining 135 subjects. Analysed subjects were found to be significantly younger (59.9 years, SD 6 vs 63.5, SD 7, p<0.0001), but there were no significant differences in gender, ethnicity and other ocular clinical parameters compared with the subjects with non-analysable images. The mean age of subjects with analysable images both at baseline and follow-up was 60.02 (baseline age, SD 6.4) years; 55.4% were women and 87.7% were Chinese (table 2). Overall, 37 had open angles by AS-OCT and 167 had at least one quadrant closed at baseline. The intraobserver and interobserver agreement (intraclass correlation) for various AS-OCT parameters measured in a subset of 30 eyes in this study were between 0.86 and 0.96, which is excellent agreement.
After 4 years, there was a significant decrease in mean AxL, mean ACD, mean ACA and mean ACV compared with baseline, and there was an increase in LV. When compared with baseline, mean values for AOD, TISA, IT and IAREA in both nasal and temporal quadrants were lower (p<0.05). A significant increase in ICURV in the temporal quadrant was also observed at follow-up (table 2).
In the univariate regression models adjusting for age and gender, shorter AxL and shallower ACD at baseline were associated with less change in TISA750 at 4 years. We categorised AxL into three groups as per the centiles (25th: 23.21 or less—51 subjects; 50th: 23.22 to 24—62 subjects; and 75th: 24.01 or more—91 subjects) and examined the age-adjusted and gender-adjusted univariate regression analysis. We found similar significant association of increased angle width narrowing, that is, mean TISA750 change in eyes with larger AxL (β=0.009, 95% CI 0.002 to 0.015, p=0.008). When compared with the lowest LV centile group (ie, 0–25th centiles, −353.3 to 233.10 µm), the highest LV centile group (ie, 75–100th centiles, 583.10–1129 µm) was significantly associated (β=−0.004, p=0.037) with less change in mean TISA750 on follow-up. Having a narrower angle at baseline was strongly associated with a lower chance of angle narrowing at follow-up, as both nasal and temporal AOD500, AOD750, TISA500 and TISA750 were associated with lower mean TISA750 change at follow-up (p<0.001). Greater mean ICURV and thinner nasal IT at 2000 µm (IT2000) were found to be associated with lower mean TISA750 change on follow-up while temporal IT2000 and IAREA were not associated with the change (table 3).
In the multivariate model, only greater mean ICURV at baseline and shorter AxL at baseline were found to be independently associated with lesser mean TISA750 change at follow-up (table 4). The model explained 14.6% of the variation in mean TISA750 change (R2=0.146).
Multivariate logistic regression evaluating the association of AS-OCT open versus closed angles at baseline to AS-OCT parameter differences at 4 years showed that baseline closed angle subjects were older, had shorter AxL, narrower angle width, smaller ACA/ACV, larger ICURV and greater LV at 4 years (data not shown).
To our knowledge, this is the first prospective study conducted on a cohort of subjects with gonioscopically open angles over time to assess changes in anterior segment parameters. Previous reports on age-related anterior segment changes have been cross-sectional.14 ,16–18 Other studies focused on the morphological changes in anterior segment parameters after laser iridotomy or phacoemulsification.19–21 In this study, we demonstrated that there was a significant decrease in ACD over 4 years in this older population, and there was a measurable decrease in ACA, ACV, as well as the angle parameters (AOD500, AOD750, TISA500 and TISA750). Increases in LV and temporal ICURV over the 4 years were also found. The apparent change in AxL could be attributed to ageing lens effect on IOLMaster readings or possible variation during inter-visit examination or both. Interestingly, Grosvenor et al22 suggested an emmetropising mechanism in adults causing true decrease in AxL with age while Leighton et al23 explained that this AxL change may contribute to shallow anterior chamber with age. However, Lee et al24 in a larger study (n=314) did not find significant AxL reduction in multivariate analysis. All these studies were cross-sectional in nature, and evidence for true change in AxL with age may require carefully conducted longitudinal studies. In multivariate analysis, greater ICURV and shorter AxL at baseline were the only independent factors associated with changes in mean TISA750 at follow-up.
Eyes with a narrower TISA750 at baseline tended to change less compared with those with wider TISA750 at baseline in this cohort. This can be explained in part by a floor effect. Similar findings were reported by Congdon et al in their study of 745 Chinese adults examined after 2 years.25 Since subjects with varying degrees of closure in AS-OCT were overselected, the magnitude of the changes found in this study may underestimate the true age-related changes that would be seen in a truly population-based cohort with open angles on gonioscopy.
With ageing, the lens probably moves anteriorly (estimated by LV) through the effect of cataract formation occupying the space in anterior chamber. The resulting iris–lens interaction forces tend to push the iris forward increasing the ICURV/pupillary block, thus obstructing the trabecular meshwork.5 ,26 Previous studies have found that iris parameters such as IT and ICURV are associated with angle closure.4 ,26 Cheung et al27 showed in their study that iris bowing (estimated by ICURV) is another critical anterior segment parameter determining angle width. We have documented prospectively that ICURV is associated with narrowing of the anterior chamber angle over time. It is to be noted that the mobile iris may be influenced over time by the systemic status of the subject at the time of image acquisition, therapy for comorbidities and so on; however, our results reflect pragmatic findings in these elderly subjects.
This study has several limitations. A major limitation of the study was the relatively low response rate of study participants after 4 years. It is possible that subjects who developed more crowded anterior segments due to cataract may have undergone cataract surgery already and would have been missed in this study. The large number of participants excluded for poor-quality AS-OCT images, ZAAP software delineation errors and indeterminate SS (39%) suggests that improvements in image acquisition as well as analysis need to be made with further studies. A previous study from our group suggested that in general SS locations could be determined in 72% of the AS-OCT images,28 which was higher than the percentage in this study (61%). This was mostly due to the fact that SS had to be clearly identified in both baseline and follow-up AS-OCT images from the same patient in order to make the analysis possible, which significantly decreased the number of subjects available for the analysis. We do not know whether eyes with unanalysable images were likely to have more or less change over time. We believe that it is unlikely that substantial bias was introduced due to this reason. Lighting condition is another confounding factor in our study result as baseline and follow-up examinations were carried out at different locations. Efforts were made to maintain similar lighting conditions and centration of the images (pupil diameter was not significantly different between baseline and follow-up, table 2), but some subtle differences may have remained. Selection of 100 out of 738 eligible subjects for the open-angle group might have introduced selection bias in recruiting more subjects with shallow AC as explained earlier. Furthermore, only static association between the risk of AC and anatomical factors was assessed in this study, other physiological factors determining AC such as iris dynamics as well as choroidal expansion/effusion were not evaluated, which partially accounts for the low R2 obtained in our multivariate regression model. Finally, the multivariate model may have been limited by the small sample size.
In conclusion, there was a significant decrease in angle width, iris area and thickness, as well as ACA/ACV, but increase in temporal ICURV and LV in this cohort of open-angle subjects followed after 4 years with AS-OCT. This decrease was found to be lower in eyes with shorter AxL and greater ICURV at baseline. This study evaluated the natural course of angle width over time and its associations with various anterior segment structures. These findings may have implications in understanding the evolution of anterior segment changes in eyes with angle closure.
YH and MB are joint first authors.
Contributors All the authors have substantially contributed in to the concept, drafting and review of the manuscript.
Funding The study was supported by grants from the SingHealth Foundation, Singapore and National Medical Research Council (NMRC), Singapore.
Competing interests None.
Patient consent Obtained.
Ethics approval Institutional review board of the Singapore Eye Research Institute and adhered to the tenets of the Declaration of HelsinkiSingHealth Centralized Institutional Review Board, Singapaore.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.