Anterior segment optical coherence tomography findings in the Infant Aphakia Treatment Study (IATS): a secondary analysis of a randomized clinical trial

Abstract

Purpose

To correlate the diagnosis of glaucoma among children in the Infant Aphakia Treatment Study (IATS) by age 10 years with anterior segment optical coherence tomography (AS-OCT) findings.

Methods

A multicenter randomized controlled trial of 114 infants with unilateral congenital cataract who were 1–6 months of age at surgery. Data on long-term glaucoma-related status and outcomes were collected when children were 10.5 years old. Participants were randomized at cataract surgery to either primary intraocular lens (IOL) or no IOL implantation (contact lens [CL]). AS-OCT findings in eyes with glaucoma were compared to eyes which did not have glaucoma and to the fellow eyes, between fellow and treated eyes, and between the IOL and CL groups.

Results

There were no significant differences in the mean nasal and temporal anterior chamber angle (ACA) or mean nasal and temporal angle opening distance (AOD) for nonglaucomatous, glaucomatous, and fellow eyes (P = 0.31, 0.16, 0.43, 0.08 resp.). There were also no significant differences in mean nasal and temporal ACA and AOD between fellow and treated eyes (P = 0.44, 0.67, 0.57, 0.38 resp.), or between IOL and CL groups (P = 0.36, 0.35, 0.49, 0.44, resp.).

Conclusions

AS-OCT confirmed that eyes with glaucoma in IATS had predominantly open angles with similar ACA and AOD to eyes without glaucoma and to fellow eyes. Furthermore, congenital cataract surgery with or without an IOL did not result in a significant difference in ACA or AOD compared to fellow eyes in IATS.

Glaucoma is a well-documented and serious complication of childhood cataract removal, with increasing frequency noted in studies with longer follow-up.^1,2 The risk of glaucoma+glaucoma suspect diagnosis after cataract removal rose from 12% (95% CI, 7%–20%) at 1 year, to 31% (95% CI, 24%–41%) at 5 years, to 40% (95% CI, 32%–50%) at 10 years in the Infant Aphakia Treatment Study (IATS).³ However the mechanism for the development of glaucoma suspect and glaucoma following congenital cataract surgery remains unclear. Most cases of glaucoma are open angle.⁴ Gonioscopy, which is the “gold standard” in evaluation of anterior chamber angles, is difficult to perform in children.⁵ Ultrasound biomicroscopy (UBM) uses ultrasonography to image the anterior segment, and studies have shown it to be consistent with gonioscopy.⁶ Nishijima and colleagues⁷ demonstrated a more narrow angle opening distance (AOD) using UBM following congenital cataract surgery compared with control eyes. The major disadvantage of UBM is that it requires an immersion shell to be placed on the eye, an approach that is generally not tolerated by unsedated young children. Due to its noninvasive nature, anterior segment optical coherence tomography (AS-OCT) has been useful for assessing the anterior segment of children with primary congenital glaucoma and juvenile open-angle glaucoma by identification of morphological features consistent with angle dysgenesis and absence of Schlemm’s canal.⁸ A pilot study of AS-OCT in 7 children following unilateral congenital cataract surgery found a larger AOD in treated eyes compared to untreated fellow eyes.⁹ In this study we report the AS-OCT findings at age 10.5 years in the IATS.

Subjects and Methods

The IATS design, eligibility criteria, surgical technique, follow-up schedule, patching and optical correction regimens, evaluation methods, and patient characteristics at baseline have been previously reported in detail.¹⁰ IATS was approved by the Institutional Review Boards of all participating centers and conformed to the requirements of the US Health Insurance Portability and Accountability Act of 1996. The treated eyes were randomized into two groups based on whether an intraocular lens (IOL) was implanted at the time of initial cataract surgery (IOL group) or the eye was left aphakic and fitted with a contact lens (CL group). Funding and approval were obtained to recall children enrolled into IATS for a detailed follow-up examination at age 10.5 years ± 3 months. In addition to visual acuity, sensorimotor, and full ophthalmic examination, measurements of axial length, optical coherence tomography (OCT) of the anterior segment and the peripapillary nerve fiber layer, and optic nerve head photography were performed. The definitions for glaucoma, glaucoma suspect, and glaucoma-related adverse events (glaucoma+glaucoma suspect) were established in the IATS, as previously reported.^3,11,12 Glaucoma was diagnosed in a study eye if the intraocular pressure (IOP) was >21 mm Hg with at least one of the following: (1) corneal enlargement, (2) asymmetrical progressive myopic shift coupled with enlargement of the corneal diameter and/or axial length, (3) increased optic nerve cupping defined as an increase of ≥ 0.2 in the cup:disk ratio, or (4) a surgical procedure was performed for IOP control.

AS-OCT was attempted on both eyes of all patients who presented for the IATS 10.5-year examination using the Heidelberg SPECTRALIS anterior segment module, ASM, which was approved by the FDA in March 2012 (Heidelberg Eye Explorer version 1.9.10.0 2014′ Heidelberg Engineering GmbH, Heidelberg Germany). AS-OCT images were to be captured prior to cycloplegic refraction per study protocol, but in some cases AS-OCT imaging was performed afterward because of the need for repeat images or the length of the study visit protocol, which also required disk photography and retinal nerve fiber layer imaging. IATS imaging specialists were trained in the use of this module, performed a readable AS-OCT on a normal patient, and used a specific protocol to perform AS-OCT on IATS patients. AS-OCT images were screened for readability prior to the patient completing the 10.5-year examination. Leung and colleauges¹³ found high reproducibility of results with the Heidelberg OCT and agreement with gonioscopy findings. An image of both the nasal and temporal angle was required to use the angle measuring software to determine the angle opening distance (AOD) and the anterior chamber angle (ACA) (eSupplement 2, available at jaapos.org and at http://www.ophthalmologyglaucoma.org).

The following protocol was used for two masked readers (AB, SF) to assess these two measurements. First, the AS-OCT image file for a particular study participant’s study eye was opened using Heidelberg software and reviewed for adequate quality, ensuring that the nasal and temporal angles of the image were both discernable. To measure the ACA and AOD, each reader used the 500 μm triangle, using the left-facing triangle for the left portion of the angle image and the right-facing triangle for the right portion of the angle image. The reader attempted to best adjust the parameters of the triangle so that it fit into the angle with the lines aligned to anatomic landmarks (eSupplement 3), and recorded the ACA and AOD on the case report form, with care to confirm the nasal versus temporal angle measurements for right versus left eyes, respectively. The process was then repeated for the fellow eye. Images in which the angle could not be evaluated (angle anatomy for both the nasal and temporal angles could not be resolved, and thus ACA and AOD could not be assessed) were judged to be inadequate (eSupplement 3, available at jaapos.org and at http://www.ophthalmologyglaucoma.org). Clinical sites were asked to retake AS-OCT images by IATS-trained image specialists on patients whose images for either the study or fellow eye were deemed inadequate.

Image quality status (no image, inadequate image, adequate image) was tabulated and the kappa statistic for agreement between fellow- and treated-eye image status calculated. For treated eyes, a two-sample t test was used to compare mean logMAR visual acuity for eyes for which images were deemed adequate versus for eyes for which there were no images or the images were deemed inadequate. Paired t tests were used to compare mean ACA and AOD measurements for the two readers. Paired t tests were also used to compare fellow and treated eye means, two-sample t tests were used to compare means for the IOL and CL groups, and analysis of variance was used to compare means for glaucomatous treated eyes, nonglaucomatous treated eyes, and fellow eyes. The two readers’ measurements were averaged for the last three analyses. Four eyes having had surgery for glaucoma were excluded from analyses other than the inter-reader comparisons and comparisons of image quality.

A significance level of 0.05 was used in all inference making; hypothesis tests and confidence intervals were two-sided. Summary statistics for logMAR visual acuity, ACA, and AOD are presented as mean with standard deviation. SAS version 9.4 (SAS Institute, Cary, NC) was used for analyses.

Results

At least one AS-OCT image was submitted for 89 of 110 subjects (81%) for central reading. However, only 54 of 110 (49%) treated eye images were adequate for measuring ACA and AOD (IOL group, 26/54 [48%]; CL group, 28/54 [52%]). Fifty-eight of 110 (53%) fellow eye images were adequate. A total of 42 subjects (38%) had adequate images for both of their eyes (IOL group, 19; CL group, 23). The kappa statistic for agreement between fellow- and treated-eye image status was 0.40 (95% CI, 0.27–0.54), indicating fair agreement.

Mean logMAR visual acuity was similar for treated eyes for which there were adequate images versus inadequate or no images (0.98 ± 0.68 vs 0.90 ± 0.59, resp. [P = 0.50]).

ACA and AOD measurements were slightly though consistently larger for reader 2. The mean paired difference between the readers’ nasal ACA measurements (reader 1 minus reader 2) was −4.2° ± 5.4° (n = 107; P < 0.001; Table 1) and between temporal ACA measurements the difference was −3.5° ± 5.5° (n = 110; P < 0.001). The mean paired difference between the readers’ nasal AOD measurements was −113.8 μm ± 177.0 μm (P < 0.001), and between temporal AOD measurements the difference was −109.1 μm ± 150.1 μm (P < 0.001).

Table 1.

Comparison of reader 1 and 2 anterior chamber angle (ACA) and angle opening distance (AOD): all eyes (treated and fellow eyes)

Study parameter	Reader 1	Reader 2	Differencea
ACA, degrees
Nasal, no.	111	107	107
Mean ± SD	35.6 ± 6.8	39.8 ± 7.1	−4.2 ± 5.4
Min, max	0, 48	0, 54	−21, 10
Temporal, no.	112	110	110
Mean ± SD	34.6 ± 8.2	38.4 ± 7.3	−3.5 ± 5.5
Min, max	0, 48	0, 51	−21, 7
AOD, μm
Nasal, no.	111	107	107
Mean ± SD	727.9 ± 168.7	841.5 ± 205.3	−113.8 ± 177.0
Min, max	0, 1094	0, 1385	−500, 691
Temporal, no.	112	110	110
Mean ± SD	701.5 ± 187.5	817.8 ± 196.8	−109.1 ± 150.1
Min, max	0, 1043	0, 1247	−520, 275

^aPaired difference (reader 1 minus reader 2). Each of the paired difference means is significantly different from 0 (P < 0.001 [paired t tests]).

All but 2 treated eyes (one with glaucoma and one without) had open angles both nasally and temporally. The mean nasal ACA of 56 fellow eyes and 47 treated eyes was 36.9° ± 5.4° vs 38.8° ± 7.6°, resp., with no significant difference between fellow and treated eyes on paired analysis of 36 children (P = 0.39; see Table 2). The mean temporal ACA of 58 fellow eyes and 49 treated eyes was 36.0° ± 5.3° vs 38.3° ± 7.2°, resp., with no significant difference on paired analysis of 37 children (P = 0.14). The mean nasal AOD of 56 fellow eyes and 47 treated eyes was 766.3 μm ± 145.2 μm vs 808.2 μm ± 193.4 μm, resp., with no significant difference on paired analysis ([P = 0.57), and the mean temporal AOD of 58 fellow eyes and 49 treated eyes was 738.9 μm ± 142.2 μm vs 809.1 μm ± 176.0 μm, resp., again with no significant difference on paired analysis (P = 0.06).

Table 2.

Anterior chamber angle (ACA) and angle opening distance (AOD) for treated and fellow eyes

Study parameter	Fellow eyesa	Treated eyesa	Paired differenceb	P valuec
ACA, degrees
Nasal, no.	56	47	36
Mean ± SD	36.9 ± 5.4	38.8 ± 7.6	−1.3 ± 8.6	0.39
Min, max	21.5, 47.5	0, 51
Temporal, no.	58	49	37
Mean ± SD	36.0 ± 5.3	38.3 ± 7.2	−1.9 ± 7.4	0.14
Min, max	20, 47.5	0, 46.5
AOD, μm
Nasal, no.	56	47	36
Mean ± SD	766.3 ± 145.2	808.2 ± 193.4	−22.4 ± 232.1	0.57
Min, max	413.5, 1106.5	0, 1239.5
Temporal, no.	58	49	37
Mean ± SD	738.9 ± 142.2	809.1 ± 176.0	−58.6 ± 180.2	0.06
Min, max	370.5, 1089.5	0, 1053

^aReader 1 and reader 2 measurements averaged for each eye.

^bAverage fellow eye measurement minus average treated eye difference.

^cP value for paired t test.

There was no significant difference in the mean nasal ACA of eyes in the CL group compared with eyes in the IOL group (P = 0.75; see Table 3) or in mean temporal ACA (P = 0.77). Similarly, there was no significant difference in the mean nasal AOD of CL eyes versus IOL eyes (P = 0.57; see Table 3), or in mean temporal AOD (P = 0.90).

Table 3.

Anterior chamber angle (ACA)^a and angle opening distance (AOD)^a for the contact lens (CL) and intraocular lens (IOL) groups

Study parameter	CL group	IOL group	P valueb
ACA, degrees
Nasal, no.	22	25
Mean ± SD	38.4 ± 9.7	39.1 ± 5.5	0.75
Min, max	0, 45.5	27.5, 51
Temporal, no.	24	25
Mean ± SD	38.0 ± 9.1	38.6 ± 5.0	0.77
Min, max	0, 46.5	25.5, 44.5
AOD, μm
Nasal, no.	22	25
Mean ± SD	790.6 ± 228.7	823.7 ± 159.5	0.57
Min, max	0, 1010	505.5, 1239.5
Temporal, no.	24	25
Mean ± SD	805.7 ± 206.5	812.4 ± 145.2	0.90
Min, max	0, 1053	488, 1019.5

^aReader 1 and reader 2 measurements averaged for each eye.

^bP value for the unequal variances two-sample t test.

There were no significant between-group differences in the mean temporal ACA for the 42 nonglaucomatous, 7 glaucomatous, and 58 fellow eyes (38.2° ± 7.7° vs 39.0° ± 4.3° vs 36.0° ± 5.3°, resp. [P = 0.16]; Figure 1A). Similarly, there were no significant between-group differences in mean nasal ACA for 40 nonglaucomatous, 7 glaucomatous, and 56 fellow eyes (39.0° ± 8.0° vs 37.3° ± 5.5° vs 36.9° ± 5.4°, resp. [P = 0.31]; Figure 1B). Mean temporal AOD measurements were not significantly different among the groups (P = 0.08; Figure 1C), nor were mean nasal AOD measurements (P = 0.43; Figure 1D).

FIG 1.

A, Distribution of temporal anterior chamber angle (ACA, averaged for two readers) by glaucoma status. Temporal mean ACA measurements did not differ significantly comparing nonglaucomatous, glaucomatous, and fellow eyes (P = 0.16 [one-way ANOVA F test]). B, Distribution of nasal anterior chamber angle (ACA, averaged for two readers) by glaucoma status. Nasal mean ACA measurements did not differ significantly comparing nonglaucomatous, glaucomatous, and fellow eyes (P = 0.31 [one-way ANOVA F test]). C, Distribution of temporal angle opening distance (AOD, averaged for two readers). Temporal mean AOD measurements did not differ significantly comparing nonglaucomatous, glaucomatous, and fellow eyes (P = 0.08 [one-way ANOVA F test]). D, Distribution of nasal angle opening distance (AOD, averaged for two readers). Nasal mean AOD measurements did not differ significantly comparing nonglaucomatous, glaucomatous, and fellow eyes (P = 0.43 [one-way ANOVA F test]).

Discussion

AS-OCT has been shown to be highly sensitive in detecting angle closure as compared to gonioscopy in adult patients.¹⁴ Spectral domain AS-OCT has also been demonstrated to obtain reproducible and consistent measurements of anterior chamber angle metrics such as ACA and AOD.¹⁵ ACA, which is simply the angle between cornea and iris at a set point, and AOD, which is the distance between the cornea and the iris along a line perpendicular to the cornea at a specified distance from the scleral spur are two of the most commonly used metrics to assess the anterior chamber angle with AS-OCT.^14,15 AS-OCT imaging in this IATS 10-year study documented that the angles of imaged children were open in the imaged region nasally and temporally in all but 2 treated eyes, one with glaucoma and the other without. The ACA and AOD in the IATS were not significantly different comparing treated and fellow eyes, IOL versus CL eyes, or glaucomatous eyes versus nonglaumatous eyes or fellow eyes. These findings are in contrast to a small study using ultrasound biomicroscopy to evaluate the anterior segment of patients following congenital cataract surgery, which demonstrated a statistically significant reduction in AOD, but not ACA, compared to normal control eyes.⁷ Nishijima and colleagues⁷ postulated that a high iris insertion was responsible for these findings, which they noted to be associated with a higher incidence of elevated IOP and glaucoma. Another study using AS-OCT of patients following congenital cataract surgery documented a smaller horizontal Schlemm’s canal diameter and cross-sectional area compared with normal controls.¹⁶ By contrast, Daniel and colleagues¹⁶ did not find a significant difference in ACA or AOD in their study, in which only the nasal anterior chamber angle structures were imaged. The AS-OCT protocol used in IATS imaged both the nasal and temporal angles simultaneously and the resolution was not sufficient to document the size of Schlemm’s canal. Furthermore the IATS imaging protocol only measured one area of the nasal and temporal angles. This protocol did not take into account anatomical changes in the superior or inferior angle, which may have contributed to the development of glaucoma.

It proved challenging to obtain readable AS-OCT images in 10-year-olds in IATS, even using experienced imaging specialists. Only 42 of 110 IATS patients (38%) had readable images of both eyes for comparison, although success with obtaining readable images of at least one eye was better. The reasons for poor image quality were not recorded in IATS, but movement artifact and poor scan resolution were the main issues noted by the two readers. These same issues were frequently noted when AS-OCT was repeated in the setting of unreadable images. The visual acuity of the treated eyes with readable AS-OCT images did not differ significantly from those without readable AS-OCT images, so poor visual acuity alone did not explain poor image quality. The inability to obtain readable AS-OCT images on the majority of IATS patients, the inability to image Schlemm's canal, and the limited number of eyes in the glaucoma group are weaknesses of this study, as is the use of cycloplegia for other study measurements before performing AS-OCT imaging in some patients, because cycloplegic agents typically cause posterior rotation of the lens/iris diaphragm and may widen angle measurements. Additionally, we lacked gonioscopic findings to compare with those of AS-OCT because gonioscopy is difficult to perform in this age group. Finally, 4 eyes that required glaucoma surgery were removed from the analyses because of the potential for postoperative angle changes from the glaucoma surgery to affect the study outcome data.

A statistically significant difference was noted between the two image readers for both ACA and AOD. Both readers are pediatric glaucoma specialists with extensive experience with gonioscopy and OCT imaging. Both image readers received training on grading AS-OCT images prior to assessing IATS study patients. All image analysis was carried out separately by each reader in masked fashion with regard to study status as glaucoma, glaucoma suspect, or nonglaucoma. A standardized protocol was utilized to assess ACA and AOD with the angle analysis tool. However, the placement of the angle tool in the angle required a manual process based on the visualized landmarks, making some variation in ACA and AOD inevitable. The two readers’ results were averaged for the main study analyses to minimize individual reader bias. However, this outcome demonstrates the limits of the software algorithm and reading protocol utilized in IATS to detect small differences in ACA and AOD.

Glaucoma following cataract surgery is one of the most common forms of childhood glaucoma in the developed world,^17,18 although the pathophysiological mechanism for the development of postoperative glaucoma remains unknown.¹⁹ Acute angle-closure glaucoma has become much less common because of advances in surgical technique that allow more complete removal of the cortex and anterior vitreous.⁷ Only 1 patient in the IATS was noted to have an acute angle-closure glaucoma event, and that patient could not be included in this AS-OCT study because phthisis bulbi developed from a retinal detachment. Cataract surgery in the first month of life has been noted to be a risk factor for the development of glaucoma.²⁰ Younger age at surgery and smaller corneal diameter, two factors that are closely related, were the only risk factors associated with the development of glaucoma or glaucoma+glaucoma suspect after 10 years’ follow-up in the IATS.³

Our AS-OCT findings in the IATS confirm that the angles of patients with glaucoma are usually open, with similar ACA and AOD compared with other treated eyes that did not develop glaucoma and compared with untreated fellow eyes. Other findings, such as a more anterior iris insertion or smaller Schlemm’s canal diameter in patients with glaucoma,^7,16 could not be confirmed in this study. Future studies with AS-OCT on patients with glaucoma following congenital cataract surgery may be aided by advances in imaging techniques that enhance rapid acquisition of high-resolution images or through the use of intraoperative AS-OCT on patients under general anesthesia to improve image quality, allowing comparison of images acquired prior to cataract surgery with those acquired at a later date.