Evaluation of fixation characteristics in amblyopia using microperimetry

To evaluate fixation characteristics in amblyopia using macular analyzer integrity assessment (MAIA) microperimetry and to investigate the factors affecting fixation stability. This prospective, cross-sectional study enrolled a total of 58 amblyopic patients who were between 8 and 55 years old. Average threshold macular sensitivity (AT) and fixation characteristics were assessed using MAIA microperimetry. Two Bivariate contour ellipse area (BCEA) fixation indices (63% and 95% proportional values) and the percentage of fixation points within 1° and 2° from the fovea (P1 and P2) were used to assess fixation stability. Non-amblyopic fellow eyes were used as the control group for comparison. AT and fixation stability indices (P1, P2, BCEA 63%, BCEA 95%) were worse in the amblyopic eyes than in the fellow eyes (p < 0.05, for all indices). There was a moderate positive correlation between best corrected visual acuity (BCVA), and AT, and P2, and a moderate negative correlation between BCVA, and BCEA indices. 48% of the eyes were eccentrically fixating (the percentage was 25% in the anisometropic group, 52% in the strabismic group, and 69% in the combined group) and 32% in the non-amblyopic eye (p = 0.052). The preferred fixation eccentricity in amblyopic eyes was significantly greater than the non-amblyopic eyes (p = 0.004), and there was a negative correlation between preferred fixation eccentricity and BCVA (p = 0.012, r = − 0.327). Our data showed a decrease in fixation stability, a positive correlation between fixation stability and BCVA, and a negative correlation between preferred fixation eccentricity and BCVA in amblyopic eyes.


Introduction
Amblyopia is reduced best-corrected visual acuity in one or less commonly two eyes caused by vision deprivation and/or abnormal binocular interaction during the first years of life [1]. It is the most common cause of unilateral preventable vision loss in children, and its prevalence has been reported to be between 1 and 3.5% [2,3]. It has several subtypes as anisometropic, strabismic, ametropic, and deprivation amblyopia. The diagnosis of amblyopia is usually made based on the best corrected visual acuity (BCVA) measurement. However, measuring visual acuity is not sufficient to determine all visual system functions [4,5]. In addition, visual functions may differ between different subtypes of amblyopia [6].
Microperimetry (MP) is a technique that can combine the morphological and functional evaluation of the macula. It provides detailed information about fixation stability and preferred retinal locus or loci (PRLs), if any, in the tested eye by tracking the eye movements [7,8]. Two Bivariate contour ellipse area (BCEA) fixation indices (63% and 95% proportional values) and the percentage of fixation points within 1° and 2° from the fovea (P1 and P2) are used to assess fixation stability. The stability of fixation is graded as stable fixation (> 75% points of fixation located in the P1), relatively stable fixation (< 75% points of fixation located in the P1, but > 75% in the P2), and unstable fixation (< 75% points of fixation located in the P2). MP also measures retinal sensitivity in the range of 0-36 dB with 37 points in a 10-degree field and gives us the average threshold macular sensitivity (AT).
Although many studies are reporting a decrease in fixation stability in the amblyopic eye [6,10,11], to our knowledge, there is no study evaluating fixation stability, fixation location, and distance of preferred retinal locus from the fovea in both amblyopic and non-amblyopic eyes in different types of amblyopia. This study aims to evaluate the fixation characteristics of amblyopic and non-amblyopic eyes in patients with anisometropic, strabismic, and combined mechanism amblyopia and to investigate the factors affecting fixation stability.

Materials and methods
This prospective, cross-sectional study was conducted at Ankara University, Faculty of Medicine, Department of Ophthalmology between July 2019 and July 2021. This study was performed in line with the principles of the Declaration of Helsinki and it was approved by the Human Research Ethics Committees of Ankara University Faculty of Medicine (Decision number: İ8-506-20, Decision date: 10.09.2020).
This study enrolled 58 amblyopic patients aged 8 years and older who were examined at Pediatric Ophthalmology and Strabismus Section. The patients were divided into three groups pure anisometropic, pure strabismic, and combined-mechanism amblyopia. All patients initially underwent a complete history taking (ocular, general, and family history), and a comprehensive eye examination, including BCVA, slit lamp biomicroscopy, and fundus examination. Distance visual acuity was tested using a Snellen chart and recorded in decimal notation. The angle of deviation was measured with an alternate prism cover test at a distance and near fixation. Binocular visual function was assessed with the Worth Four-Dot test at a distance and near. Stereopsis was tested with the Titmus stereopsis test. Cycloplegic retinoscopy was performed for the detection of refractive errors and converted to spherical equivalent (SE). Amblyopia was defined as BCVA < 20/25 in the affected eye and two or more lines of visual acuity difference between the two eyes. In the presence of at least ≥ 1.0 D hyperopic difference, ≥ 3.0 D myopic difference, or ≥ 1.5 D astigmatic difference between the two eyes and absence of strabismus, we considered this condition as pure anisometropic amblyopia.
The average threshold macular sensitivity and fixation analysis were performed with the MAIA microperimetry (CenterVue, Padova, Italy) instrument. The MP expert mode 4-2 strategy was used in the study. BCEA fixation indices (63% and 95%) and P1 and P2 parameters as percentiles were used to assess fixation stability. An MP output of the non-amblyopic and the amblyopic eye of a patient with strabismic amblyopia is shown in Fig. 1. The location of fixation was evaluated as either central (central 4°) or eccentric fixation. Fixation location was expressed as superior, inferior, temporal, or nasal according to its position relative to the fovea center [24] (Fig. 2). All MPs were performed by a single technician and fixation classification was assessed by a single researcher (D.A.).
Age, BCVA, refractive error, angle of deviation, binocular vision, stereopsis, macular sensitivity, fixation stability, and fixation location were evaluated in this study. The data from amblyopic eyes were compared with those from non-amblyopic fellow eyes. The eyes with nystagmus; eyes with BCVA below 0.05 or above 0.8; eyes with less than two Snellen lines between amblyopic and non-amblyopic eyes; eyes that failed to complete the MP test; or patients with the concomitant ophthalmologic or systemic disease were excluded from the study.
Continuous variables were reported as mean ± (SD) and were compared using Student's t-test or Mann-Whitney U tests based on the normality of data. Distribution pattern of the continuous variables was determined by Shapiro Wilk test. Categorical variables were reported as n (%). Chi-Square test was used to compare categorical variables. Kruskal Wallis test or one-way ANOVA analysis was performed for comparison of variables between different amblyopia subtypes according to the distribution pattern. Tukey test was used as the post-hoc test for ANOVA analysis. Bonferroni correction was applied if required after Kruskal Wallis test. The correlation analysis between BVCA and MP test results was evaluated by Pearson correlation or Spearman correlation according to the normality of data. Two-sided p-values of less than 0·05 were considered statistically significant. Statistical analyses were performed using SPSS version 22.

Results
A total of 58 patients (30 female and 28 male) were included in the study. The mean age was 21 ± 11 years (ranging between 8 and 55 years). The amblyopia subtype was anisometropic amblyopia in 16 cases (27.6%), strabismic amblyopia in 29 cases (50%), and combined mechanism amblyopia in 13 cases (22.4%). The amblyopic eye was the right eye in 48.3% of the patients. The mean SE was 0.89 ± 2.83 D in the amblyopic eyes, and 0.62 ± 2.46 D in the non-amblyopic eyes. Suppression was present in 75.9% and 69% of the cases with the Worth Four-Dot test at distant and near, respectively. There was no gross stereopsis in 12 patients. The mean stereopsis was 536 ± 844 s of arc for the remaining cases. Descriptive features are given in Table 1.
The mean BCVA was 0.38 ± 0.24 in amblyopic eyes. (0.38 ± 0.25 in the anisometropic group; 0.41 ± 0.26 in the strabismic group; 0.34 ± 0.20 in the combined group). In the anisometropic group, the most common refractive errors were hypermetropia (43.8%), and compound hypermetropic astigmatism (31.2%), and the mean SE was 2.55 ± 2.88 D. The frequency of esotropia and exotropia was similar in the strabismic group and the combined group (44.8%).
The mean AT was 25.5 ± 4.8 dB in the amblyopic eyes and 27.7 ± 2.8 dB in the non-amblyopic eyes. The mean AT in amblyopic eyes was worse than in non-amblyopic eyes (p = 0.003). The mean P1 value was 58 ± 31, P2 value was 80 ± 21 in amblyopic eyes, while the P1 value was 74 ± 25, P2 value was 89 ± 13 in non-amblyopic eyes. Fixation stability was lower in amblyopic eyes than in non-amblyopic eyes (p = 0.002 for P1; p = 0.036 for P2). The comparison of the MP test results of amblyopic and non-amblyopic eyes is shown in Table 2.
According to intergroup comparisons among different amblyopia subtypes, there was no difference among groups in age, gender, BCVA, and binocular visual functions (p = 0.057, p = 0.227, p = 0.8, p = 0.818). The median stereopsis was significantly better in the anisometropic group than that in the combined group (excluding patients without stereopsis by Titmus test) (p = 0.045). There was no difference among groups in AT, P1, and P2 (p = 0.231, 0.101, and 0.16, respectively). There was no correlation between retinal sensitivity as AT and fixation stability parameters (P1, P2, BCEA 63%, BCEA 95%) and age, gender, refractive error type, SE, binocular visual functions, presence of stereopsis, and level of stereopsis. There was a moderate positive correlation between BCVA and AT and between BCVA and P2, and a moderate negative correlation between BCVA and BCEA 63% and between BCVA and BCEA 95%. The correlation coefficients and p values are shown in Table 3.   The mean distance between the preferred fixation points and the fovea in the amblyopic eyes was significantly greater than in the non-amblyopic eyes (p = 0.004). The mean distance in the amblyopic eyes was 2.97° ± 2.56°, while it was 1.83° ± 1.37° in the non-amblyopic eyes. There was a negative correlation between the distance of the preferred fixation point to the fovea and BCVA in the amblyopic eyes. There was no correlation in the non-amblyopic eyes. The correlation coefficients and p values are shown in Table 3. The frequency of eccentric fixation was 48% among the amblyopic eyes and 32% among the non-amblyopic eyes, and there was no statistically significant difference between the amblyopic and the non-amblyopic eyes (p = 0.114). Eccentric fixation was found to be 25% in the anisometropic group, 52% in the strabismic group, and 69% in the combined group (p = 0.052). The fixation location was in the nasal quadrant in 31% of cases with esotropia and 26% of cases with exotropia. In amblyopia, the BCVA was 0.45 ± 0.23 in central fixation and 0.31 ± 0.24 in eccentric fixation. In amblyopia, the mean BCVA in eyes with eccentric fixation was worse than in eyes with central fixation (p = 0.016). There was no difference between the amblyopia groups for the preferred fixation eccentricity and fixation location in both amblyopic eyes and the non-amblyopic eyes (p = 0.08 for the preferred fixation eccentricity in the amblyopic eyes, p = 0.052 for the fixation location in the amblyopic eyes; p = 0.978 for the preferred fixation eccentricity in the non-amblyopic eyes, p = 0.713 for the fixation location in the non-amblyopic eyes).

Discussion
In our study, AT and fixation stability in the amblyopic eyes were found to be significantly worse than in the non-amblyopic eyes. There was no correlation between fixation stability parameters (P1, P2, BCEA 63%, BCEA 95%) and age, gender, refractive error type, SE, binocular visual functions, presence of stereopsis, and level of stereopsis. There was a moderate positive correlation between BCVA with AT and P2 and a moderate negative correlation between BCVA with BCEA 63%, and BCEA 95%. The frequency of eccentric fixation was higher in the strabismic and combined amblyopia groups than in the anisometropic amblyopia group. The preferred fixation eccentricity was greater in amblyopic eyes than the non-amblyopic eyes, and there was a negative correlation between preferred fixation eccentricity and BCVA.
A decrease in fixation stability of the amblyopic eye [5-7, 10, 11, 13, 14, 19, 21-23] and a positive correlation between BCVA and fixation stability have been reported in many studies [7,10,12,15,19,25]. Similarly, in our study, fixation stability was found to decrease in the amblyopic eye, and it was found to be significantly worse than in the non-amblyopic eyes. Significant correlations were found between fixation stability parameters (P2, BCEA 63%, BCEA 95%) and BCVA. It was determined that the fixation stability increased as the BCVA increased. Tarutta et al. [12] found a strong correlation between fixation stability and BCVA in strabismic and refractive amblyopia. Similarly, Chung et al. [25] reported a significant correlation between visual acuity and fixation stability in anisometropic, strabismic, and combined-mechanism amblyopia. Subramanian et al. [7] detected a decrease in fixation stability in patients with amblyopia. However, while they found significant correlations between BCVA and fixation stability in the strabismic group and combined-mechanism amblyopia group, they could not find a significant correlation between BCVA and fixation stability in the pure anisometropic amblyopia group. Dikkaya et al. [5], similar to this study, did not find any relationship between BCVA and fixation stability in their study, which included only patients with anisometropic amblyopia. It was thought that this result might be related to the more frequent occurrence of extraocular motor changes and extrafoveal-unstable fixation in strabismic amblyopia than in anisometropic amblyopia [26]. Unlike our study, Dickman et al. [10] reported that they did not find a significant difference in fixation stability between the amblyopic eye and the non-amblyopic fellow eye. This result was thought to be related to the fact that the BCVA values of the amblyopic eyes in their study (between 20/50 to 20/32) were better than the BCVA values in our study (between 20/400 to 20/25) in which 28 (48.3%) eyes had a BCVA even worse than 20/50.
Many studies have reported reduced macular sensitivity of amblyopic eyes. Central, paracentral, and generalized depression can be observed [7,10,[18][19][20]. In addition, scotoma was reported in non-amblyopic eyes using a scanning laser ophthalmoscope by Johnson [19]. Dickman et al. [10] found small, focal, central scotoma and generalized depression in both the strabismic and anisometropic amblyopia groups and found the macular sensitivity of non-amblyopic eyes to be better than the amblyopic eyes. Similarly, macular sensitivity in the amblyopic eyes was significantly lower than in the non-amblyopic fellow eyes in our study. A positive correlation between the decrease in RNFL thickness and macular sensitivity in amblyopic eyes has been suggested in previous studies [10,21]. Therefore, the decrease in fixation stability in amblyopic eyes was thought to be related to the decrease in retinal sensitivity in the macula.
Although the basic mechanisms are similar in strabismus and anisometropic amblyopia, they have differences [27,28]. Strabismus affects visual development earlier in life than anisometropia and may negatively affect interneuron transmission and neuron development [7]. Especially in strabismic amblyopia, central scotoma may sometimes occur in the strabismic eye, and eccentric fixation may develop due to suppressive mechanisms during the critical visual development period [7,29]. In our study, 48% of amblyopic eyes had eccentric fixation, and 32% of non-amblyopic eyes had eccentric fixation. There was no difference in fixation location between the amblyopic and non-amblyopic eyes. Eccentric fixation was found to be 25% in the anisometropic amblyopia group, 52% in the strabismic type, and 69% in combined mechanism amblyopia. Similar to the literature, the frequency of eccentric fixation was higher in the strabismic and combined groups in our study. The median stereopsis in our study was better in the anisometropic amblyopia group than in the combinedmechanism amblyopia group. It was thought that this might be related to the lower frequency of eccentric fixation in the anisometropia group.
Strabismus was present in approximately 72% of our cases, and esotropia and exotropia were equal in frequency in this group. Eccentric fixation was observed in 55% of amblyopic eyes with strabismus. The fixation location was in the nasal quadrant in 31% of cases with esotropia and 26% of cases with exotropia. No study was found in the literature on fixation location. Only Bangerter [30], reported that the localization of extrafoveal fixation is usually nasal in esotropia and temporal in exotropia. In our study, it was not found that the fixation location changed depending on the strabismus type.
The distance of fixation to the fovea affects fixation stability. It has been reported that the fixation ability is significantly worse in eccentric fixation compared to central fixation in patients with amblyopia and the preferred fixation eccentricity is very effective on fixation instability [16]. This entity is closely related to BCVA [24]. In our study, the preferred fixation eccentricity was greater in amblyopic eyes than in non-amblyopic eyes, and there was a negative correlation between the preferred fixation eccentricity and BCVA. This condition was attributed to the fact that the highest resolution and sensitivity in the retina is in the fovea. Chung et al. [31] reported that the further away from the fovea, the lower the resolution in the retina.
Subramanian et al. [7] reported that there is a relationship between stereopsis and fixation instability in the amblyopic eye. Our study revealed no correlation between fixation stability parameters and binocular visual functions, presence of stereopsis, and level of stereopsis. This result might be related to the exclusion of 12 patients without stereopsis from the statistical analysis and limiting the number of our cases. The coexistence of strabismus in most of the patients with amblyopia in our study may have affected fixation stability, visual acuity, fusion, and stereopsis from different aspects, causing extrafoveal fixation to be more frequent.
It has been thought that a more stable fixation may change monocular visual acuity and binocular vision in amblyopic eyes. Some studies showed that the increase in visual acuity in the amblyopic eye was accompanied by an increase in fixation stability after treatment [7,[32][33][34][35]. However, it is still controversial whether visual acuity affects fixation or fixation stability affects vision [12]. In clinical practice, it was reported that the evaluation of fixational eye movements should be questioned in patients with amblyopia that do not respond to classical amblyopia treatments such as glasses, occlusion, and atropine [33].
Acoustic BFT performed with MP has been shown to improve fixation stability in patients with low vision with central scotoma [9,36]. It has been reported in the literature that knowing the fixation characteristics of amblyopia may be beneficial in treatment and follow-up [12]. Maneschg et al. [11] reported that after surgical alignment of strabismus in adult amblyopic eyes, fixation stability can be improved following BFT. The study of Lapajne et al. [37], one of the limited number of studies conducted, showed that vision training with visual evoked potentials (VEP) biofeedback after the critical period (8-17 years) is also beneficial in amblyopia. Esposito et al. [38] reported improvement in visual function with biofeedback training in young people with anisometropic amblyopia. It was stated that the hardware and software required for modern biofeedback training to improve fixation stability in amblyopic cases could be collected in a small device such as a smartphone that amblyopic patients can use at home [30].
Our study evaluated functional changes in amblyopic eyes, but their relationship with retinal structural changes and contrast sensitivity was not examined. It would make the study more powerful by examining the correlation between the structural changes (such as macular thickness, RNFL, foveal volume, and choroidal thickness) in both the amblyopic eye and the non-amblyopic eye and the microperimetry parameters, visual acuity, stereopsis, and contrast sensitivity level with additional tests. Our study included the non-amblyopic eyes as the comparison group, but it would be better to include a control group consisting of completely healthy eyes. Another limitation of the study is the number of cases. Statistical subgroup analyses to be performed with a more extended number of participants will strengthen the study. This study differs from its counterparts in the literature in evaluating the fixation stability, fixation location, and preferred fixation eccentricity using microperimetry in amblyopia, both in the amblyopic eyes and in the non-amblyopic eyes, and in investigating the correlation between fixation location and BCVA.
In conclusion, our study shows that macular sensitivity and fixation stability are decreased in eyes with amblyopia, and there is a moderate positive correlation between fixation stability and visual acuity, and a negative correlation between preferred fixation eccentricity and visual acuity in these eyes. Monitoring retinal functions and fixation characteristics with microperimetry in patients with amblyopia will help us to understand the cause and mechanism of low vision.
Author contributions All mentioned authors and collaborators have made a substantial contribution and have approved the final version.
Funding No external funding.

Conflict of interest
The authors indicate no financial support or financial conflict of interest. The authors have indicated they have no financial relationships with any company.
Ethical approval This study was performed in line with the principles of the Declaration of Helsinki and it was approved by the Human Research Ethics Committees of Ankara University Faculty of Medicine (Decision number: İ8-506-20, Decision date: 10.09.2020).