Relationship Between Suppression Scotomas and Stereoacuity in Anisometropic Amblyopia With Successfully Treated Visual Acuity

Abstract

Purpose: The purpose of this study was to investigate the relationship among suppression scotoma size, stereoacuity, and four-prism base-out test (4ΔBOT) results in anisometropic amblyopia with successfully treated visual acuity.

Methods: We included 103 cases of anisometropic amblyopia successfully treated for visual acuity without strabismus. Stereoacuity was measured using a Randot Stereotest. The size of the suppression scotomas was measured using a new device, the polarized four dot (P4D) test. This is a modification of the Worth 4 dot test (W4D) device. The patients were divided into three groups based on the 4ΔBOT results: normal (group A = 29 cases), subnormal (group B = 48 cases), and abnormal (group C = 26 cases) response groups. The horizontal diameter of the suppression scotomas and stereoacuity in logarithmic values with a base of 20 seconds of arc (″) were compared among the 3 groups.

Results: The mean age at P4D testing was 8.4 ± 2.1 years. The average horizontal diameters of the suppression scotomas were 0.35 ± 0.79Δ, 2.01 ± 0.82Δ, and 5.50 ± 2.72Δ in groups A, B, and C, respectively, showing significant differences (A versus B: P < 0.0001, A versus C: P < 0.0001, and B versus C: P < 0.0001; 1-way ANOVA). The average logarithmic stereoacuity were 1.07 (24.95″), 1.22 (38.84″), and 1.47 (82.79″) in groups A, B, and C, respectively, thereby showing significant differences between the groups (A versus B: P < 0.0001, A versus C: P < 0.0001, and B versus C: P < 0.0001; 1-way ANOVA). Stereoacuity and horizontal diameter of the suppression scotoma were strongly correlated (r = 0.732, P < 0.0001).

Conclusions: The suppression scotoma size measured using P4D correlated significantly with stereoacuity and the 4ΔBOT results.

Anisometropia is defined as the relative refractive difference between the eyes. It is a causative factor of amblyopia and abnormal binocularity. Anisometropic amblyopia may occur partly because of the effect of image blur on the development of the visual system in the affected eye and partly because of suppression and interocular conflict to combine different image blurs.¹ Higher anisometropia is associated with greater risk and greater severity of amblyopia.^2–4 Despite the successful improvement in visual acuity with fully cycloplegic refractive correction and occasional patching of the dominant eye, patients with anisometropic amblyopia rarely achieve normal stereoacuity of <60 seconds of arc (″).^5–7 Most patients with anisometropic amblyopia and abnormal binocularity have a suppression scotoma. Suppression scotomas may be associated with stereoacuity. However, little is known about the objective relationship between suppression scotoma and stereoacuity in patients with anisometropic amblyopia.^8–14 Brooks et al. reported the effects of experimentally induced anisometropia on stereoacuity in healthy adults.¹⁵ They found that the size of suppression scotoma increased proportionally with anisometropia and that stereoacuity was similarly degraded by the induced anisometropia. Other reports have also reported the relationship among the degree of experimentally induced anisometropia, stereoacuity, and suppression scotoma in healthy adults.^16–23 However, these reports did not discuss the relationship between the size or depth of suppression scotoma and stereoacuity in real patients. 

The exact mechanism by which anisometropia leads to a marked decrease in stereoacuity remains unknown. 

The Worth four-dot test (W4D) and the four-prism base-out test (4ΔBOT) are methods used to detect suppression scotoma. However, these methods are only qualitative. Trials have been reported on the quantitative measurement of suppression scotomas, including dichoptic global motion coherence and binocular contrast summation.^{12,13,16,24–26} Although these methods are rigorous, their protocols are usually difficult to apply to children because they require complicated devices and numerous trials. Therefore, the success of these trials in children remains limited. Few reports are available regarding trials on the quantitative measurement of suppression scotomas with protocols that may be applied to children.^27,28 However, the number of cases in children younger than 6 years of age are at most 20.^27,28 Thus, the clinical significance of scotoma suppression remains poorly understood in children. 

The purpose of this study was to measure suppression scotomas using the polarized four-dot (P4D) test, which was jointly developed by our co-authors Teiji Yagasaki and Nihon Tenganyaku Kenkyusho Co., (Nagoya, Aichi, Japan; Fig. 1a, b). These data would be used to determine the association among the size of suppression scotomas, stereoacuity, and 4ΔBOT results among 103 children with anisometropic amblyopia. 

Figure 1.

 

(a) Photograph of the polarized four-dot test. (b) Photograph of the polarized glasses. (c) Vision of amblyopic and fellow eye when the patient was wearing polarized glasses.

Methods

There were 103 patients with anisometropic amblyopia (55 boys and 48 girls), who visited our hospitals between 1995 and 2021. Patients who had a hyperopic refractive error difference between the eyes of >1.50 diopters (D) or an astigmatic refractive error difference of >1.50 D with successfully treated visual acuity without manifest strabismus were included. Patients with accommodative esotropia, ocular disorders other than anisometropic amblyopia, neurologic disorders, cognitive impairment, or systemic diseases were excluded. Amblyopia was defined as the difference in the best-corrected visual acuity (BCVA) of worse than 0.0 log MAR in the affected eye and 0.2 log MAR difference between eyes. The detailed characteristics of all cases are mentioned in Tables 1a and b. 

Table 1A.

 

Details of the Characteristics in Cases of Hypermetropic Anisometropia

 

Table 1B.

 

Details of the Characteristics in Cases of Astigmatic Anisometropia

 

All patients were prescribed fully corrected glasses based on cycloplegic refraction. They were ordered to occlude the sound eye 2 hours daily if no fusional response occurred with Bagolini striated glasses. All patients underwent refractive adaptation before any occlusion. Amblyopia was resolved by achieving a BCVA of 0.0 log MAR or better in the previously amblyopic eye and less than 0.2 log MAR difference between eyes, and stabilization of this BCVA for 6 months or longer was considered as successful treatment. 

Suppression scotomas were measured using P4D (see Fig. 1a), a device which is a modification of the W4D. In P4D, 2 different polarization filters are attached to the device and glasses to perform binocular separation instead of the color filters used in the W4D device. This is because, in W4D, binocular color rivalry may be caused by looking through green and red filters with the left and right eyes, respectively. 

There were 4 dots in front of the P4D device. The diameter of each dot was 4 cm. The distance from the middle of a dot to the center of the device was 8 cm. Dot A in Figure 1a is nonpolarized. Dots B and C were polarized in the vertical direction, and dot D was polarized in the horizontal direction, using Wratten-type linear polarizing filters. 

The patient wore polarizing filters in the form of glasses over their refractive correction lenses (see Fig. 1b), which was corresponded to the same polarizing filters as the filters to the P4D device. The polarizing filter in the horizontal direction was corresponded to the glasses on the side of the amblyopic eye, and that on the side of the fellow eye was corresponded to the polarizing filter in the vertical direction (Fig. 1c). Wearing glasses, the patient could see three dots (A, B, and C) with the fellow eye and two dots (A and D) with the amblyopic eye (see Fig. 1c). Therefore, during a definite fusion response in both eyes of the patient, an individual could see four dots (A, B, C, and D). However, when suppression occurs in the amblyopic eye of the patient, one could only see three dots (A, B, and C). 

The P4D device was set at 30 cm from the patient. The visual angle in prism diopters (PD) is 8 cm/0.3 m (i.e. 26.7 PD). The visual angle between the two opposite dots changes with changes in the distance between the patient and the device. The patient was instructed to wear the glasses, look at four dots on the device, and tell the examiner when the number of dots changed from four to three. The P4D device was moved slowly away from the patient until the four dots (i.e. a definite fusion response) changed to three dots (i.e. suppression in the amblyopic eye). The visual angle at the point where the dots changed from four to three corresponded to the suppression scotoma radius (i.e. the radius of the suppression size in PD, which is 8 cm/change distance in meters; Fig. 2a). As shown in Figure 2a, we can measure the lower radius of the suppression scotoma. The dots can be rotated by 360 degrees every 90 degrees and can be set to four position patterns (patterns ①, ②, ③, and ④ in Fig. 2b). This enables the dot that the patient cannot see when the amblyopic eye is under suppression (i.e. see dot D in Fig. 1a) to be placed in four different locations (see patterns ①, ②, ③, and ④ in Fig. 2b). We can measure the upper radius in pattern ①, the right radius in pattern ②, the lower radius in pattern ③, and the left radius in pattern ④. We added right and left radii (i.e. the horizontal diameter of the suppression scotoma) and used them to calculate the effective suppression scotoma size in the anisometropic eye. The clinician measuring the size of the suppression scotoma was blind to the 4ΔBOT result. This quantitative measurement of the suppression scotoma was repeated three consecutive times in one patient, and the average was used for statistical analysis. To evaluate the reproducibility of this method of quantitative measurement of the suppression scotoma using P4D, we compared each of the 3 horizontal diameters of the suppression scotomas in 30 consecutive patients tested most recently from 103 patients included in this study. The intraclass correlation coefficient of this method was also evaluated. 

Figure 2.

 

(a) Quantitative measurement of a suppression scotoma. (b) The dots of P4D can be rotated 360 degrees every 90 degrees and can be set to four position patterns. It enables the measurement of upper, lower, right, and left radii of suppression scotoma, respectively.

Stereoacuity was evaluated using a Randot Stereotest (Stereo Optical Co., Chicago, IL, USA). As a mathematical rule, all stereoacuity values were converted to logarithmic values with a base of 20″, which is the minimum value of the Randot Stereotest, such as 20″ (1), 30″ (1.14), 40″ (1.23), 50″ (1.31), 70″ (1.42), 100″ (1.54), 140″ (1.65), 200″ (1.77), 400″ (2), and 500″ (2.07). Patients with no detectable stereoacuity (i.e. nil) were assigned a value of 7000″ (2.96). We investigated the correlations between logarithmic stereoacuity and horizontal diameter of suppression scotomas in all patients. 

All patients were examined using the 4ΔBOT, a partial quantitative test used to detect the existence of a suppression scotoma larger than 4Δ. In this test, the patient was instructed to watch a fixation target. The examiner placed a 4Δ base-out prism in front of the patient's eye and observed the responses of both eyes. For a person who did not have a >4Δ suppression scotoma, the image was shifted by the prism and the eye had an inward movement. 

Based on Herring's law of equal innervation, the fellow eye moves outward (phase 1). The patient simultaneously experienced diplopia and the fellow eye had a fusional convergence movement. This eye stopped at the convergence position (phase 2). The test was used to examine both the eyes of all patients several times. If the eye had a suppression scotoma of >4Δ when the examiner placed a 4Δ base-out prism in front of the patient's eye, the eye did not have an inward movement, and there were no other ocular responses (i.e. “no movement”). If the fellow eye had suppression scotoma, it had no phase 2 response (i.e. “no reverse”). Patients who demonstrated in phases 1 and 2 were defined as “normal responders.” Patients who often had a normal response but occasionally had no reverse response were defined as “subnormal responders.” All other responses were defined as “abnormal responders.” The clinician conducting the 4ΔBOT was blind to the stereopsis findings and size of the suppression scotoma in all cases. Based on these results, all patients were divided into 3 groups: normal responders, subnormal responders, and abnormal responders, comprising 29, 48, and 26 patients, respectively. 

Statistical analyses were performed using the IBM SPSS Statistics for Windows, version 29.0 (IBM Corp., Armonk, NY, USA) and Microsoft Excel for Windows, version 16 (Microsoft Corp., Redmond, WA, USA). The horizontal diameter of the suppression scotoma, stereoacuity, pretreatment log MAR visual acuity, degree of anisometropia, and patient's age at the first visit to our hospital were compared among the three groups using 1-way ANOVA. The Bonferroni correction was applied to compare the three groups. The relationships between the size of the suppression scotomas and stereoacuity were investigated using Pearson's correlation coefficient. The significance level was adjusted to a family wise error of P < 0.017 for multiple comparisons. The tests were considered statistically significant when P < 0.05. We also evaluated the estimation accuracy for each linear relationship between the logarithmic stereoacuity and suppression scotoma size of the three groups using the root mean square error (RMSE) based on the statistical proof by SATISTA, Co., Ltd. (Uji, Kyoto, Japan). 

This study was approved by the ethics committee of the Japan Community Health Care Organization Chukyo Hospital (Nagoya, Japan). Informed consent was obtained from all parents or guardians of the patients. All medical procedures were performed in accordance with the tenets of the Declaration of Helsinki. 

Results

The Reproducibility of Quantitative Measurement of the Suppression Scotoma Using P4D

To evaluate the reproducibility of this method of quantitative measurement of suppression scotomas using P4D, we compared each measurement of the horizontal diameters of the suppression scotomas. Each measurement was obtained 3 times per patient using the P4D in 30 representative cases from the 103 cases in this study. The averages of the horizontal diameters of the suppression scotoma in the representative 30 cases were 4.40 ± 2.42Δ at the first time, 4.09 ± 2.27Δ at the second time, and 3.93 ± 1.98Δ at the third time. The intraclass correlation coefficient was excellent (r = 0.929). Therefore, the reliability is high when the size of the suppression scotoma is measured thrice for each of the 30 representative cases using the P4D device. 

Comparison Between the Three Groups Based on the 4ΔBOT Results

Altogether, 103 patients who met the inclusion criteria were identified and included in the study. The mean age of patients at their first visit to our hospital was 4.7 ± 1.5 years (range = 1–9 years). The mean age of the patients when they were tested with P4D was 8.4 ± 2.1 years (range = 4–15 years). 

The average horizontal diameters of the suppression scotoma were 5.50 ± 2.72Δ (1.64–9.99Δ), 2.01 ± 0.82Δ (0–5.39Δ), and 0.35 ± 0.79Δ (0–2.59Δ) in the abnormal, subnormal, and normal groups, respectively. The intergroup differences in suppression scotoma extents were significant (normal versus subnormal = P < 0.0001, normal versus abnormal = P < 0.0001, and subnormal versus abnormal = P < 0.0001; 1-way ANOVA; Fig. 3a). 

Figure 3.

 

(a) The averages of the sizes of the suppression scotomas in the three groups. (b) The averages of stereoacuity in the three groups.

The average logarithmic stereoacuity, based on 20″’, were 1.07 (24.95″), 1.22 (38.84″), and 1.47 (82.79″) in the normal, subnormal, and abnormal groups, respectively. The intergroup differences were also significant (normal versus subnormal = P < 0.0001; normal versus abnormal = P < 0.0001; and normal versus abnormal = P < 0.0001; 1-way ANOVA; Fig. 3b). 

No significant differences existed among the three groups in age at the first visit, pretreatment visual acuity, or degree of anisometropia (Table 2). The correlation between logarithmic stereoacuity and horizontal diameter of suppression scotoma was strong in all 103 patients (r = 0.732, P < 0.00001; Fig. 4a). 

Table 2.

 

Average of Pretreatment Log MAR Visual Acuity, Degree of Anisometropia, and Age at the First Visit at our Hospital

 

Figure 4.

 

(a) Correlation between stereoacuity and the size of a suppression scotoma. (b) The linear estimation in the three groups and in the total cases.

When we performed a linear estimate between the logarithmic stereoacuity (y) and suppression scotoma size (x) for the three groups, we found no significant difference in their slopes (normal group = 0.047 [95% confidence interval = −0.021 to 0.114]; subnormal group = 0.028 [95% confidence interval = −0.023 to 0.078]; abnormal group = 0.046 [95% confidence interval = 0.025 to 0.067], P = 0.797). Therefore, we considered that the slopes of X1 to Y2 in each group were approximately similar. We obtained one linear estimating equation for all three groups. As a result, the overall linear estimating equation obtained was Y2 = 1.091 + 0.064 × X1, with an R-squared value of 0.536 (Fig. 4b). When the estimation accuracy in each group was evaluated using RMSE, no significant difference was observed when estimating with the overall equation compared to the RMSE when estimating each group individually. This result indicates the validity of using one linear estimating equation to predict for each group (RMSE: group 1 = 0.095 vs. 0.104 [estimation in each group versus estimation in total]; group 2 = 0.130 vs. 0.134 [estimation in each group versus estimation in total]; and group 3 = 0.187 vs. 0.196 [estimation in each group versus estimation in total]). 

Discussion

Stereoacuity and the size of suppression scotoma were strongly correlated. The size of suppression scotoma was smallest in the normal group, followed by (in increasing order) the subnormal and abnormal groups. Stereoacuity was superior in the normal group, followed by (in decreasing order) the subnormal and abnormal groups. The 4ΔBOT results were closely associated with the size of suppression scotoma and the stereoacuity in anisometropic amblyopia. 

Lysons et al.²⁹ divided 532 anisometropic children (4–5 years old) without manifest strabismus into 4 groups based on visual outcome. They evaluated the relationship between visual acuity after treatment and the results of the 4ΔBOT. All 324 children in the 2 groups in which the children had achieved equal visual acuity in the affected eye, the fellow eye, after maximum optical treatment until 2 months or 6 months of treatment, had a normal response to 4ΔBOTs. However, all 208 children in the other 2 groups, wherein the children did not achieve equal visual acuity after maximum optical treatment and/or occlusion therapy until the sixth month of treatment, had an abnormal response to 4ΔBOT. This finding suggests that the visual outcome after treatment is associated with 4ΔBOT results in patients with anisometropic amblyopia. In our study, we chose children with anisometropic amblyopia whose visual acuity was better than 0.0 log MAR, and differences in visual acuity between eyes were less than 0.2 log MAR after treatment without manifest strabismus. In our study, 74 of 103 patients had a subnormal or abnormal response to 4ΔBOT, even if their visual acuity was better than 0.0 log MAR. Furthermore, the findings of our study suggest that 4ΔBOT results are closely associated with stereoacuity in anisometropic amblyopia. 

Narasimhan et al. measured the relative depth of a suppression scotoma between both eyes in 19 anisometropic amblyopia and 14 strabismic amblyopia and 6 aniso-amblyopia children (age = 5–16 years) using dichoptic global motion coherence.¹² They investigated the relationship between visual acuity and the depth of suppression scotoma. They reported that when the suppression was deeper, the visual acuity was worse. Similarly, the results of our study showed that the larger the suppression scotoma, the worse tend to be the pretreatment visual acuity although no statistically significant differences were observed. 

Moreover, they reported that the depth of suppression was milder in anisometropic amblyopia than in strabismic amblyopia, and that the depth of suppression was milder in patients with normal binocular response than in those with abnormal binocular response. Agrawal et al.¹⁰ and Li et al.¹³ also reported similar results. Similarly, in our present study, the size of the suppression scotoma was milder in the patients of anisometropia with normal response than abnormal response in 4ΔBOT. 

Then, what is the meaning of existence of a suppression scotoma? Livingstone et al. suggested that suppression scotoma may be a secondary response to avoid diplopia and visual confusion in strabismus and/or amblyopia.³⁰ However, in our study, all patients achieved 0.0 log MAR and had no manifest strabismus. Thus, this hypothesis cannot explain the significance of suppression scotoma in the patients in our study. 

Visual information travels from the retina to the V1 area in the occipital lobe, mostly along the parvocellular (P pathway) and magnocellular (M pathway) pathways. The P pathway is more sensitive to low temporal and high spatial frequency stimuli, and the M pathway is more sensitive to high temporal and low spatial frequency stimuli.^31–37 Although amblyopia is mainly an abnormality of the P pathway,^31,33 some authors also reported temporal deficits in amblyopia, including anisometropic amblyopia.^38–42 The axon of horizontal neurons is thought to be 7 mm long, corresponding to approximately 5Δ in the visual field around the fovea. Ocular dominance columns of the right and left eyes connect in the V1 area with a pair of horizontal neurons and process the amount of stereoscopic parallax.³⁶ A suppression scotoma <10Δ may corresponds to the axon length of a pair of horizontal neurons from the retina to the visual cortex V1.³⁶ Therefore, the size of the suppression scotoma may reflect a P pathway abnormality at the stage of the development of visual function and may correspond to the minimum threshold of the stereoscopic parallax processed in the visual cortex. This is only speculation derived from the reports of animal experiments. However, we think that it is important to speculate about the origin of the suppression scotoma from those reports. 

In this study, a certain number of patients in the subnormal and abnormal groups were diagnosed with “microtropia with identity.” This diagnosis was made based on the results of 4ΔBOT, as these patients were too young to assess eccentric fixation using a fixation ophthalmoscope.⁴³ Patients in the normal group had been diagnosed with “anisometropic amblyopia without microtropia.” In this study, when we performed a linear estimate between the logarithmic stereoacuity and the size of the suppression scotoma in the three groups, the slopes of the lines of the three groups were not significantly different. This finding suggests that all three groups have identical spectra. Thus, anisometropic amblyopia and microtropia may be different aspects of the identical spectrum disorder with suppression scotoma and stereoacuity worse than 60″. 

Available reports showed no studies in which the patients were younger than those in the present study (Table 3).^{10–14,27,28,44} The quantitative measurement of suppression scotomas using P4D is one of the valid methods for children. 

Table 3.

 

Reports About Trials for Quantitative Measurement of Suppression Scotomas

 

Our study has a few limitations. The P4D and 4ΔBOT were tested at only one time point and were not followed up several times over the course of time. Garretty reported moderate agreement between 4ΔBOT and microperimetry, which are the most commonly used methods for diagnosing microtropia with identity.⁴³ Garretty reported that the results of the microtropia test may change over time.⁴³ In the report, she stated the importance of not relying on only one test at one time point and that patients should be reassessed as the treatment progresses.⁴³ Houston et al.⁴⁵ also reported that the suppression area can be altered by occlusion therapy. Moreover, the size of a suppression scotoma may change before and after maximum optical treatment and occlusion therapy. Therefore, future studies should assess the size of a suppression scotoma several times, especially before and after treatment. Furthermore, when this abnormality occurs, the occurrence of suppression scotoma remains unclear. More studies are required regarding this aspect of scotoma suppression. 

In conclusion, the size of the suppression scotoma was closely associated with the results of 4ΔBOT and stereoacuity. The findings of our study suggest that scotoma suppression plays an important role in anisometropic amblyopia and microtropia. However, there were no differences in the relationship between the size of suppression scotoma and stereoacuity in the three groups, where patients in the present study were divided by responses of 4ΔBOT. These results suggest that microtropia cannot be distinguished from anisometropic amblyopia after successful treatment of visual acuity in the affected eye using the 4ΔBOT. 

Acknowledgments

The authors would like to thank SATISTA Co., Ltd. (Uji, Kyoto, Japan) for statistical analysis and Editage (www.editage.com) for English language editing. 

Disclosure: K. Hozumi, None; T. Yagasaki, None; Y. Yokoyama, None; A. Yagasaki, None; Y. Haga, None; R. Eboshita, None