- Email: [email protected]
Risk factors for abnormal binocular vision after successful alignment of accommodative esotropia
Risk Factors for Abnormal Binocular Vision After Successful Alignment of Accommodative Esotropia Sherry L. Fawcett, PhDa,b and Eileen E. Birch, PhDa,b Purpose: The purpose of this study was to identify clinical factors associated with abnormal binocular vision outcomes among children with accommodative esotropia (ET) whose eyes were successfully realigned with spectacles only or with spectacles and surgery. Methods: The participants were 69 children with accommodative ET who were followed up prospectively from the time of diagnosis. Clinical factors examined in this study included high accommodative convergence–to–accommodation (AC/A) relationship, high hyperopia, anisometropia, age of onset, and duration of eye misalignment. Binocular vision was assessed using measures of stereopsis, fusional vergence, sensory foveal fusion, and motion visual-evoked potential (mVEP). Results: Children with a high AC/A relationship are 2.2 times more likely to have an absence of fusional vergence than are children with a normal AC/A relationship. Children having a duration of constant eye misalignment ⱖ 4 months before being successfully treated are 4.6 times more likely to have abnormal stereopsis, 33 times more likely to have no stereopsis, 37 times more likely to have an absence of fusional vergence, 31 times more likely to have an absence of sensory foveal fusion, and 17 times more likely to have an asymmetric mVEP response than children with a duration of constant ET diagnosed at 0 to 3 months.Conclusions: Following successful eye alignment, as many as 75% of patients with accommodative ET had anomalous binocular vision. A high AC/A relationship poses a significant risk for abnormal fusional vergence only. A constant eye misalignment lasting ⱖ 4 months poses a significant risk for anomalous binocular vision on all measures studied. (J AAPOS 2003;7:256 –262) ccommodative esotropia (ET) is a convergent misalignment of the eyes that typically develops between 18 and 48 months of age. It is associated with activation of the accommodation reflex and, frequently, with moderate to severe hyperopia.1 Although many children with accommodative ET are effectively treated with lens correction, the esodeviation of a substantial number of children cannot be treated effectively by spectacles alone. After being aggressively treated with spectacles, monocular occlusion, prisms, and/or surgery, some patients never achieve fully normal binocular vision even when the esodeviation is fully corrected. Given the relatively late onset of accommodative ET, why do some children with accommodative ET have abnormal binocular vision after treatment? One hypothesis is
A
From the Retina Foundation of the Southwest,a and the Department of Ophthalmology,b University of Texas Southwestern Medical Center, Dallas, Texas. Supported by National Institutes of Health Grant No. EY05236. Presented in part at the annual meeting of The Association of Research in Vision and Ophthalmology, Fort Lauderdale, FL, April 30-May 5, 2000. Submitted November 19, 2002. Revisions accepted April 8, 2003. Reprint requests: Sherry Fawcett, PhD, Retina Foundation of the Southwest, 9900 N Central Expressway, Suite 400, Dallas, TX 75231. Copyright © 2003 by the American Association for Pediatric Ophthalmology and Strabismus. 1091-8531/2003/$35.00 ⫹ 0 doi:10.1016/S1091-8531(03)00111-3
256
August 2003
that prolonged eye misalignment results in a permanent disruption of the mechanisms underlying foveal fusion. This hypothesis is supported by a recent finding that early treatment of accommodative ET (eg, spectacle correction while ET is still intermittent or within 3 months of the development of a constant ET) helps to prevent a permanent loss of fine stereopsis.2 The association between various clinical factors and binocular vision outcomes among patients with accommodative ET has been the topic of several published articles.3-8 Retrospective medical chart reviews have shown an association between a high accommodative convergence to accommodation (AC/A) relationship and abnormal binocular vision and deterioration among patients with accommodative ET. Patients with accommodative ET and a high AC/A ratio have a lower incidence of foveal fusion and fine stereopsis compared with patients having normal AC/A ratios.3-6 No studies published to date have examined the relationship between hyperopia or anisometropia and binocular sensory outcomes of patients with accommodative ET. However, a recent study examining the role of anisometropia in the development of accommodative ET by patients with hyperopia indirectly supports a hypothesis that anisometropia may be a significant risk factor for abnormal binocular vision among patients with accommodative ET. Among the patients who developed accommodative ET in this study, those with anisometropia had an Journal of AAPOS
Journal of AAPOS Volume 7 Number 4 August 2003
increased risk for unsatisfactory alignment with spectacles alone.7,8 Because the duration of constant eye misalignment impacts binocular sensory outcomes among patients with accommodative ET,2,6 patients with anisometropia who develop esotropia may be at increased risk for abnormal sensory outcomes because the need for surgery increases the time to achieve orthotropia. The purpose of the present study was to identify clinical factors associated with abnormal binocular vision outcomes among children with accommodative ET. Clinical factors examined in this study included high AC/A relationship, high hyperopia, anisometropia, age at onset, and duration of constant eye misalignment. Binocular vision was assessed using measures of stereopsis, fusional vergence, sensory foveal fusion, and presence or absence of a nasotemporal mVEP asymmetry.
MATERIAL AND METHODS Participants Sixty-nine children with diagnosed accommodative ET were referred by local pediatric ophthalmologists to the Pediatric Laboratory at the Retina Foundation of the Southwest for sensory evaluation. The diagnosis of accommodative ET was made because the esotropia manifested by each of these patients was initially completely corrected with spectacles. At follow-up examinations, however, some patients’ esotropia became manifest once more and was no longer correctable using hypermetropic corrections alone, thus requiring eye muscle surgery. These children were enrolled in a prospective research study investigating the development and maintenance of binocular vision. Age at onset of esotropia in the children ranged between 2 months and 5 years (mean, 22 months). Eye realignment was achieved at 10 months to 8 years with optical correction alone (N ⫽ 46) or with optical correction and surgery (N ⫽ 23). Final sensory testing for this study was completed when the patients were between 3 and 12 years of age (mean, 4.5 years). Inclusion in this study required children to have achieved and maintained good eye alignment (ⱕ 8 PD) with spectacles or a combination of spectacles and surgery for at least 6 months at the time of testing. Exclusion criteria included neurological abnormalities, coexisting systemic disease, previous ocular surgery, preterm birth, and difference in best-corrected visual acuity between the eyes greater than two lines. Informed consent was obtained from parents before participation. The research protocol was approved by the Institutional Review Board of the University of Texas Southwestern Medical Center and was performed according to the tenets of the Declaration of Helsinki. Measures of Binocular Vision Random Dot Stereopsis. Random dot stereoacuity was measured using the Randot Butterfly Test (Stereo Optical, Chicago, IL); the Randot (Version 2) Shapes Test (Stereo Optical); the Randot Preschool Stereoacuity Test
Fawcett and Birch
257
(Stereo Optical); and the Lang 1 Test (Fresnel Prism and Lens, Scottsdale, AZ). Coarse stereopsis was measured using the Titmus Fly Test (Stereo Optical). All tests were administered according to the protocol standardized by Pediatric Eye Disease Investigator Group (PEDIG) for multicenter clinical trials. Any children who were unable to cooperate with shape-based tests were assessed using the Infant Random Dot Stereoacuity Cards (Stereo Optical).9 A 2AFC preferential looking technique using a two down– one up staircase format beginning with 435 seconds of arc disparity was used to quantify stereoacuity when the Infant Random Dot Stereoacuity Cards were used. The session concluded when eight staircase reversals were obtained; when the child made at least six errors at the largest disparity; or when the first six trials were correct at the smallest disparity. Stereoacuity threshold was determined as the average disparity of the last six reversals.9 The normality of individual patient’s stereoacuity scores was determined using normative values derived using these tests.9,10 On the Randot Preschool Stereoacuity Test, 95% of children age ⱖ 3 years scored ⱖ 200 seconds of arc, and 95% of children age ⱖ 4 years scored ⱖ 100 seconds of arc.10 On the Infant Random Dot Stereoacuity Cards, 95% of children scored ⱖ 150 seconds of arc by 24 months of age.9 For this study, stereoacuity values greater than 95% tolerance limits for the test administered were considered to be abnormal. Motor Fusion/Fusional Vergence. Motor fusion/fusional vergence was measured using the 4 PD base-out test. A 4 PD loose prism was applied base out in front of one eye while the patient was instructed to fixate a small toy held approximately 33 cm out. Sensory foveal fusion was considered to be present in patients who produced a fusional vergence eye movement by the fellow eye in response to placement of the prism in front of the test eye. Grading for the 4 PD base-out test was pass/fail. To minimize the occurrence of false-negative results, passing required the ability to elicit a fusional vergence eye movement response in both eyes. Sensory Fusion. Sensory foveal fusion was measured using the Worth 4-dot test at 3 m (0.66°) according to the PEDIG standardized protocol. Grading of the Worth 4-dot test was pass/fail. Passing required reporting four dots. Failure required reporting two, three, or five dots. Only children who were able to cooperate or complete testing were included in the analysis of sensory foveal fusion (N ⫽ 50). Motion Visual Evoked Potential (mVEP). Nasotemporal mVEP asymmetry is associated with an absence of foveal fusion in patients with infantile ET and accommodative ET.11 In the current study, mVEP is included as an additional objective measure of binocularity. The NuDiva sweep-VEP system was used to acquire and analyze the mVEPs.12 Motion VEPs were elicited by jittering 1 cm/degree sine wave vertical gratings separated by 90° spatial phase at 6 Hz (12 reversals/sec). Gratings
258
Fawcett and Birch
were displayed on a 34° ⫻ 25° field using a high-resolution video monitor with a mean luminance of 161.6 candela/m2 and a contrast of 80%. Participants viewed the stimuli monocularly from a distance of 50 cm. Monocular viewing was achieved by covering one of the eyes with an opaque patch. During a single 10-second trial, mVEPs were recorded when the corneal reflection of the video monitor was centered in the child’s pupil. Recording was interrupted when fixation was lost. Fixation of the stimulus was maintained by dangling a small toy immediately in front of the monitor. The electroencephalogram (EEG) was recorded from two bipolar derivations (01 and 02), 2 to 2.5 cm to the left and right of a common reference electrode (0z) on the midline. A ground electrode was placed 2.5 cm above the reference electrode. The potential differences were amplified (gain ⫽ 10,000; ⫺3db cutoff at 1 and 100 Hz). The EEG was adaptively filtered (bandpass) in real time (sampling time ⫽ 397 Hz) to isolate the VEP. The EEG was subjected to Fourier analysis to extract the amplitude and phase of the VEP at the first (F1) and second (F2) harmonics of the stimulus. All analyses were based on vector averages of five or more trials for each eye. Amplitude, phase, signal-to-noise ratio, and a T 2 circ statistic were computed based on the vector average of the individual trials for each eye.13 The amplitude at the response frequency was compared with the average amplitude recorded at the noise frequencies adjacent to the recording frequency. A response was considered to be significant if its signal-to-noise ratio was ⬎ 3:1. Symmetric and asymmetric mVEPs produce characteristic Fourier spectra. A symmetric mVEP yields a Fourier spectra composed primarily of F2, where there is a peak of activity at twice the temporal frequency of the stimulus, ie, equal responses to the onset of nasal and temporalward horizontal motion. An asymmetric mVEP yields a Fourier spectra composed primarily of F1, where there is a peak of activity at the fundamental frequency of the stimulus, ie, a stronger response to the onset of motion in one direction of motion than the other. The presence of motion asymmetry was determined using the presence versus absence of a “bowtie.”11,12 By this method, motion asymmetry is distinguished by a significant F1 response component in the same channel(s) for both eyes with an interocular phase differences of 180 ⫾ 40°. Risk Factors Examined High AC/A Relationship. AC/A relationship was classified, using the heterophoria method, as normal or high according to the difference between the distance (6 m) and near (33 cm) prism and alternate cover measurements of eye alignment. Measures of distance and near eye alignment were measured using an accommodative target with hyperopic correction. A dichotomous classification of high AC/A was given for any patients with ⬎ 10 PD difference
Journal of AAPOS Volume 7 Number 4 August 2003
between the distance and near measurements of eye alignment and normal AC/A for any patients with ⱕ 10 PD difference between the distance and near measurements of eye alignment. Note that the heterophoria method, although commonly used in the clinic, does not distinguish between high AC/A ratio and nonaccommodative convergence excess.14 The range of difference between the distance and near measurements of eye alignment was between 0 and 45 PD. Twenty patients were classified with a high AC/A and 49 patients were classified with a normal AC/A ratio. High Hyperopia. Hyperopia was considered moderate to high if it was ⱖ 4.50 D spherical equivalent. Thirtyseven patients had hyperopia ⱖ 4.50 D, and 32 patients had hyperopia ⬍ 4.50 D. Anisometropia. Anisometropia was defined as ⱖ 1 D spherical equivalent refractive error difference between the two eyes. Fifteen patients had anisometropia, and 54 patients had no anisometropia. Age of Onset. Onset age was based on the parents’ and ophthalmologist’s reports. For the purpose of analysis, onset age was grouped as 0 to 6 months (congenital) versus ⱖ 7 months (acquired) and 0 to 17 months (infantile) versus ⱖ 18 months (late onset). Ten patients had onset at 0 to 6 months, and 21 patients had onset at 0 to 17 months. Duration of Eye Misalignment. Duration of eye misalignment was calculated as the difference between the onset age of constant eye misalignment and patient age at eye realignment. For the purpose of analysis, duration of constant eye misalignment was grouped as 0 to 3 months versus ⱖ 4 months and 0 to 11 months versus ⱖ 12 months. Twenty-three patients had a constant eye misalignment lasting 0 to 3 months, and 35 patients had a constant eye misalignment lasting 0 to 11 months. Statistics Risk Factor Analysis. The relationship between each risk factor and abnormal binocular vision outcome was calculated via the Mantel-Haenszel odds ratio (OR): OR ⫽ [P1/(1-P1)]/[P0/(1-P0)], where P1 is the incidence of the outcome in patients exposed to the risk factor, and P0 is the incidence of the outcome in patients not exposed to the risk factor. Because the incidence of abnormal binocular vision outcomes in this study population exceeded 10%, the Mantel-Haenszel odds ratios were corrected to relative risks (RR):15,16 RR ⫽ OR/(1-P0) ⫹ (P0 ⫻ OR) Relative risk data—along with 95% confidence intervals (CI)—are presented in Figures 1 and 3 through 6. Factors with a 95% CI lower limit ⬎ 1.0 pose a significant risk for an anomalous binocular vision outcome (equivalent to P ⬍ .05). Because the RR and its associated 95% CI cannot be computed when the observed incidence of an adverse out-
Journal of AAPOS Volume 7 Number 4 August 2003
come is equal to zero, for the purpose of analyses, any zeros were converted to 0.5 to provide a conservative estimate of the RR. Secondary Analysis. When a factor was found to pose a significant risk for abnormal stereoacuity, a secondary analysis was conducted to compare median stereoacuities of subgroups via the Kruskal-Wallis one-way analysis of variance (ANOVA) on ranks. When a factor was found to pose a significant risk for abnormal fusion (4 PD base-out test, Worth 4-dot test, or mVEP test), a secondary analysis was conducted to determine the significance of proportions of patients within subgroups either passing or failing.
RESULTS Stereoacuity scores ranged from 40 seconds of arc to nil. Binocular vision results are summarized in the Table 1. Because binocular vision is not an all-or-none sensory function, it is common to have different results on measures of sensory fusion, motor fusion, and stereopsis. It is believed that different tests measure different degrees of sensory and/or motor fusion and that different tests tap different underlying neural mechanisms of binocularity. The RR for each of the risk factors for abnormal stereopsis are summarized in Figure 1. The only clinical factor studied that posed a significant risk for abnormal stereopsis was duration of constant eye misalignment. Children with eye misalignment lasting ⱖ 4 months were 4.6 times more likely to have abnormal stereopsis than those with constant eye misalignment lasting 0 to 3 months. Kruskal-Wallis one-way ANOVA on ranks examining the difference between median stereoacuity scores by patients with 0 to 3 months versus 4 to 11 months versus ⱖ 12 months of constant eye misalignment showed a significant effect of duration on stereoacuity outcome (H ⫽ 54.373, P ⬍ .001). As shown in Figure 2, significant differences between the median stereoacuity scores were found between patients with 0 to 3 months versus 4 to 11 months and 0 to 3 months versus ⱖ 12 months of eye misalignment. The RRs of each of the risk factors for nil stereopsis are summarized in Figure 3. Duration of constant eye misalignment was the only clinical factor studied that posed a significant risk for nil stereopsis. Children with eye misalignment lasting ⱖ 4 months were 33 times more likely to have nil stereopsis than were children with 0 to 3 months of constant eye misalignment. There was a trend toward a higher proportion of patients with nil stereopsis in the group with 4 to 11 months of constant eye misalignment versus the group with 0 to 3 months of constant eye misalignment (25% versus 0%, z ⫽ 1.87, P ⫽ .06). Significant differences were found between the proportions of patients with nil stereopsis with 0 to 3 months versus ⱖ12 months of constant eye misalignment (0% versus 85%, z ⫽ 6.05, P ⬍ .001) and between 4 to 11 months versus ⱖ 12 months of constant eye misalignment (25% versus 85%, z ⫽ 3.54, P ⬍ .001).
Fawcett and Birch
259
TABLE 1. Incidence of abnormal binocular vision Random dot stereoacuity (secs arc) 40-60 100-250 9 16 (8 normal)*
400-3000 12
nil 32
Motor fusion/fusional vergence (4 PD base-out prism test) Pass Fail 32 37 Sensory fusion (Worth 4-dot test at 3 m) Pass Fail 27 23 Motion VEP Symmetric 49
Unable 19
Asymmetric 20
*Two hundred seconds of arc is within the normal range for 3-year-old children, and 100 seconds of arc is within the normal range for 4-year-old children; VEP, visual-evoked potential.
FIG 1. Relative risks for abnormal stereopsis as a function of high AC/A ratio (⬎10 D difference between near and distance eye alignment); hyperopia (ⱖ 3 D); anisometropia (ⱖ 1 D spherical equivalent refractive error difference between the two eyes); age at onset of constant ET (0 to 6, 0 to 17, and 0 to 24 months); and duration of constant ET (ⱖ 4 and ⱖ 12 months). The 95% CIs for relative risk are shown as horizontal bars. The relative risk for each of the clinical factors is shown as a vertical line within each horizontal bar. Significant risk is observed with 95% CI lower limit ⱖ 1.0.
The RR associated with each of the clinical factors for an absence of fusional vergence is shown in Figure 4. Both high AC/A ratio and duration of constant eye misalignment posed significant risks for failing the 4 PD base-out test. Children with a high AC/A ratio were 2.2 times more likely to fail the 4 PD base-out test than were children with a normal AC/A ratio. Children with eye misalignment lasting ⱖ 4 months were 37 times more likely to fail the 4 PD base-out test than were children with 0 to 3 months of constant eye misalignment. The difference between the proportions of patients who failed the 4 PD base-out test with 0 to 3 months versus 4 to 11 months of constant eye misalignment was statistically significant (0% versus 50%, z ⫽ 3.25, P ⬍ .001). Also significant were the differences in 4 PD base-out test failure (1) between the patients with 0 to 3 months versus ⱖ 12 months of constant eye misalignment (0% versus 88%, z ⫽ 6.26, P ⬍ .001) and (2)
260
Fawcett and Birch
Journal of AAPOS Volume 7 Number 4 August 2003
FIG 4. Relative risks for abnormal fusional vergence (failing the 4 PD base-out test) as a function of each of the risk factors. Other details as in Figure 1. FIG 2. Random dot stereoacuity (seconds of arc) as a function of duration of constant ET. Shaded areas indicate the interquartile range (25th and 75th percentile), and error bars indicate 10th and 90th percentiles for each duration group. The asterisks indicate a significant difference of median stereoacuity compared with the 0- to 3-month group as determined by Dunn’s pairwise comparisons (P ⬍ .05). The horizontal lines in the shaded boxes represent the median stereoacuity values (100 and 1775 seconds of arc). The median stereoacuity value for patients with ⱖ12 months of constant ET is nil.
FIG 5. Relative risks for absence of foveal fusion (failing the Worth 4-dot test at 3 months) as a function of each of the risk factors. Other details as in Figure 1.
FIG 3. Relative risks for nil stereopsis as a function of each of the risk factors. Other details as in Figure 1.
between 4 to 11 months versus ⱖ 12 months of constant eye misalignment (50% versus 88%, z ⫽ 2.33, P ⫽ .02). The RR associated with each of the clinical factors for an absence of sensory foveal fusion subsequent to successful eye realignment is shown in Figure 5. Duration of constant eye misalignment is the only clinical factor studied that posed a significant risk for failing the Worth 4-dot test at 3 m. Children with eye misalignment lasting ⱖ 4 months were 31 times more likely to fail the Worth 4-dot test at 3 m than were children with 0 to 3 months of constant eye misalignment. The difference between the proportions of patients who failed the Worth 4-dot test with 0 to 3 months versus 4 to 11 months of constant eye misalignment was statistically significant (0% versus 37.5%, z ⫽ 2.22, P ⬍ .03). Also significant were the differences in Worth 4-dot test failure (1) between the patients with 0 to 3 months versus ⱖ 12 months of con-
FIG 6. Relative risks for mVEP asymmetry as a function of each of the risk factors. Other details as in Figure 1.
stant eye misalignment (0% versus 91%, z ⫽ 5.59, P ⬍ .001) and (2) between 4 to 11 months versus ⱖ 12 months of constant eye misalignment (37.5% versus 91%, z ⫽ 2.58, P ⫽ .01). The RR associated with each of the clinical factors for an asymmetric mVEP subsequent to successful eye realignment is shown in Figure 6. Duration of constant eye misalignment was the only clinical factor studied that posed a significant risk for an asymmetric mVEP. Children with eye misalignment lasting ⱖ 4 months were 17 times more likely to have an asymmetric mVEP than children with 0 to 3 months of constant eye misalignment. The
Journal of AAPOS Volume 7 Number 4 August 2003
difference between the proportions of patients who exhibited an asymmetric mVEP with 0 to 3 months versus 4 to 11 months of constant eye misalignment was not statistically significant (0% versus 9%, z ⫽ 0.35, P ⫽ .73). The difference between the proportions of patients who exhibited an asymmetric mVEP with 0 to 3 months versus ⱖ 12 months of constant eye misalignment (0% versus 47%, z ⫽ 3.52, P ⬍ .001) was significant, and a similar trend was found when comparing patients with 4 to 11 months versus ⱖ 12 months of constant eye misalignment (9% versus 47%, z ⫽ 2.58, P ⫽ .055).
DISCUSSION In this study, we examined the RR of high AC/A relationship, high hyperopia, anisometropia, age at onset, and duration of constant eye misalignment on binocular vision outcomes in nonamblyopic children with accommodative ET, age 3 to 12 years, after successful treatment. Binocular vision was assessed using four measures: stereopsis, motor fusion/fusional vergence, sensory foveal fusion, and mVEP response. A high AC/A relationship was found to pose a significant risk only for an absence of fusional vergence as measured by the 4 PD base-out test. Within this sample of patients, a high AC/A relationship was not found to pose a significant risk for anomalous binocularity on other measures of binocularity including stereopsis, sensory foveal fusion, and the symmetry of the mVEP response. The lack of association between high AC/A relationship and abnormal stereopsis is in agreement with an earlier study.2 Both studies showed that patients with high AC/A ratios can achieve fine stereopsis with early bifocal spectacle and/or surgical intervention. Conversely, it is important to note that the cohort studied excluded any patients who had not achieved eye alignment within 8 PD at the time of testing. Had we included patients who had not been successfully aligned, the prevalence of nil stereopsis would have been much higher. Because the prevalence of poor alignment may be higher among children with high AC/A ratios,3-6 we may have underestimated the overall risk of high AC/A relationship on poor binocular sensory outcomes in the general clinical population. High hyperopia was not found to pose a significant risk for abnormal binocular vision among patients with accommodative ET. In our sample of patients, hyperopia ranged between 0.25 and 8 D, with an equal proportion of patients falling within the categories of mild hyperopia (44%) and moderate to high hyperopia (56%). Within these categories, there was an equal distribution of children with abnormal versus normal outcomes on each of the measures of binocular vision. Anisometropia was not found to pose a significant risk for abnormal binocular vision in patients with accommodative ET. The small representation of patients with anisometropia within this sample (22%), however, decreases our confidence in making any strong conclusions about
Fawcett and Birch
261
this finding. A larger sample of patients with accommodative ET who have anisometropia is necessary to make a reliable calculation of the RR of anisometropia for abnormal binocular vision. Age at onset was not found to pose a significant risk for abnormal binocular vision for patients with accommodative ET whose eyes were successfully realigned by the time of testing. Infantile-onset versus childhood-onset accommodative ET was found to pose no greater risk for an abnormal binocular vision outcome after successful eye realignment. A delay of surgical alignment of infantile ET longer than 3 months results in poor stereoacuity outcomes.17 A similar observation has been made in a cohort of children with accommodative ET.2 Here we show that other measures of binocular sensory function (fusional vergence, sensory foveal fusion, and mVEP response) share a similar duration of abnormal eye alignment for disruption. Prolonged eye misalignment can be the result of a number of factors including delayed medical referrals from primary health care providers, third-party payer requirements for second opinions, poor compliance with treatment, and an initially ineffective course of treatment. An important consideration in treatment of strabismus is understanding the importance of prompt eye alignment to rescue binocular visual function. To expand our understanding of the time course of the critical period of development, continued susceptibility, and recovery of binocular vision, we examined the differences of binocular sensory function in children with accommodative ET who experienced 0 to 3 months of eye misalignment, 4 to 11 months of eye misalignment, and longer than 12 months of eye misalignment before being successfully treated. Significant differences between the stereoacuity scores of successfully realigned patients with 0 to 3 months (100 seconds of arc) versus 4 to 11 months (1775 seconds of arc) of constant ET indicate that with increasing duration of eye misalignment, the amount of stereoacuity that can be achieved after successful treatment decreases. An analysis of the differences between the proportions of patients with no measurable stereopsis who have 0 to 3 months (0%) versus 4 to 11 months (25%) versus ⱖ 12 months (85%) of constant eye misalignment indicates that if not moderate stereopsis, at least coarse stereopsis can be achieved when treatment is effective in correcting eye alignment after up to 11 months of constant eye misalignment. Fusional vergence, sensory foveal fusion, and symmetric mVEP responses are equally, if not more susceptible to loss, as is stereopsis. An analysis of the differences between the proportions of patients with no fusional vergence, no sensory foveal fusion, and asymmetric mVEPs, respectively, who have 0 to 3 months (0%, 0%, and 0%) versus 4 to 11 months (50%, 37.5%, and 9%) versus ⱖ 12 months (88%, 91%, and 47%) of constant eye misalignment indicates that with increasing duration of eye misalignment, fusional vergence, sensory foveal fusion, and symmetric mVEP responses are lost.
262
Fawcett and Birch
Overall, the data presented here support the hypothesis that the duration of constant eye misalignment is the primary risk factor for adverse binocular vision outcomes after successful treatment of children with accommodative ET. References 1. Parks MM. Ocular motility and strabismus. Hagerstown (MD): Harper and Row; 1975. p. 99-105. 2. Fawcett SF, Leffler JN, Birch EE. Factors influencing stereoacuity in accommodative esotropia. J AAPOS 2000;4:15-20. 3. Parks MM. Management of acquired esotropia. Br J Ophthalmol 1974;58:240-7. 4. Pratt-Johnson JA, Tilson G. Sensory outcome with nonsurgical management of esotropia with convergence excess (a high accommodative convergence/accommodation ratio). Can J Ophthalmol 1984;19: 220-3. 5. Ludwig IH, Parks MM, Getson PR, Kammerman LA. Rate of deterioration in accommodative esotropia correlated to the AC/A relationship. J Pediatr Ophthalmol Strabismus 1988;25:8-12. 6. Wilson ME, Bluestein EC, Parks MM. Binocularity in accommodative esotropia. J Pediatr Ophthalmol Strabismus 1993;30:233-6. 7. Weakley DR, Birch EE. The role of anisometropia in the development of accommodative esotropia. Trans Am Ophthalmol Soc 2000; 98:71-9.
Journal of AAPOS Volume 7 Number 4 August 2003 8. Weakley DR, Birch EE, Kip K. The role of anisometropia in the development of accommodative esotropia. J AAPOS 2001;5:153-7. 9. Birch EE, Solomao SR. Infant random dot stereocards. J Pediatr Ophthalmol Strabismus 1998;35:86-90. 10. Birch EE, Williams C, Hurtes J, Lapa MC, ALSPAC. Children in Focus study team. Random dot stereoacuity of preschool children. J Pediatr Ophthalmol Strabismus 1997;34:217-22. 11. Fawcett SF, Birch EE. Motion VEPs, stereopsis, and bifoveal fusion in children with strabismus. Invest Ophthalmol Vis Sci 2000;41: 411-6. 12. Norcia AM, Garcia H, Humphrey R, Holmes A, Hamer RD, OrelBixler D. Anomalous motion VEPs in infants and in infantile esotropia. Invest Ophthalmol Vis Sci 1991;32:436-9. 13. Victor JD, Mast J. A new statistic for steady state evoked potentials. Electroencephalogr Clin Neurophysiol 1991;78:378-88. 14. von Noorden GK, Campos EC. Binocular vision and ocular motility. 6th ed. St. Louis: Mosby; 2002. p. 85-100. 15. Zang J, Yu KF. What’s the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA 1998;280: 1690-1. 16. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22: 719-48. 17. Birch EE, Fawcett S, Stager DR. Why does early surgical alignment improve stereoacuity outcomes in infantile esotropia? J AAPOS 2000;4:10-14.
A
From the Retina Foundation of the Southwest,a and the Department of Ophthalmology,b University of Texas Southwestern Medical Center, Dallas, Texas. Supported by National Institutes of Health Grant No. EY05236. Presented in part at the annual meeting of The Association of Research in Vision and Ophthalmology, Fort Lauderdale, FL, April 30-May 5, 2000. Submitted November 19, 2002. Revisions accepted April 8, 2003. Reprint requests: Sherry Fawcett, PhD, Retina Foundation of the Southwest, 9900 N Central Expressway, Suite 400, Dallas, TX 75231. Copyright © 2003 by the American Association for Pediatric Ophthalmology and Strabismus. 1091-8531/2003/$35.00 ⫹ 0 doi:10.1016/S1091-8531(03)00111-3
256
August 2003
that prolonged eye misalignment results in a permanent disruption of the mechanisms underlying foveal fusion. This hypothesis is supported by a recent finding that early treatment of accommodative ET (eg, spectacle correction while ET is still intermittent or within 3 months of the development of a constant ET) helps to prevent a permanent loss of fine stereopsis.2 The association between various clinical factors and binocular vision outcomes among patients with accommodative ET has been the topic of several published articles.3-8 Retrospective medical chart reviews have shown an association between a high accommodative convergence to accommodation (AC/A) relationship and abnormal binocular vision and deterioration among patients with accommodative ET. Patients with accommodative ET and a high AC/A ratio have a lower incidence of foveal fusion and fine stereopsis compared with patients having normal AC/A ratios.3-6 No studies published to date have examined the relationship between hyperopia or anisometropia and binocular sensory outcomes of patients with accommodative ET. However, a recent study examining the role of anisometropia in the development of accommodative ET by patients with hyperopia indirectly supports a hypothesis that anisometropia may be a significant risk factor for abnormal binocular vision among patients with accommodative ET. Among the patients who developed accommodative ET in this study, those with anisometropia had an Journal of AAPOS
Journal of AAPOS Volume 7 Number 4 August 2003
increased risk for unsatisfactory alignment with spectacles alone.7,8 Because the duration of constant eye misalignment impacts binocular sensory outcomes among patients with accommodative ET,2,6 patients with anisometropia who develop esotropia may be at increased risk for abnormal sensory outcomes because the need for surgery increases the time to achieve orthotropia. The purpose of the present study was to identify clinical factors associated with abnormal binocular vision outcomes among children with accommodative ET. Clinical factors examined in this study included high AC/A relationship, high hyperopia, anisometropia, age at onset, and duration of constant eye misalignment. Binocular vision was assessed using measures of stereopsis, fusional vergence, sensory foveal fusion, and presence or absence of a nasotemporal mVEP asymmetry.
MATERIAL AND METHODS Participants Sixty-nine children with diagnosed accommodative ET were referred by local pediatric ophthalmologists to the Pediatric Laboratory at the Retina Foundation of the Southwest for sensory evaluation. The diagnosis of accommodative ET was made because the esotropia manifested by each of these patients was initially completely corrected with spectacles. At follow-up examinations, however, some patients’ esotropia became manifest once more and was no longer correctable using hypermetropic corrections alone, thus requiring eye muscle surgery. These children were enrolled in a prospective research study investigating the development and maintenance of binocular vision. Age at onset of esotropia in the children ranged between 2 months and 5 years (mean, 22 months). Eye realignment was achieved at 10 months to 8 years with optical correction alone (N ⫽ 46) or with optical correction and surgery (N ⫽ 23). Final sensory testing for this study was completed when the patients were between 3 and 12 years of age (mean, 4.5 years). Inclusion in this study required children to have achieved and maintained good eye alignment (ⱕ 8 PD) with spectacles or a combination of spectacles and surgery for at least 6 months at the time of testing. Exclusion criteria included neurological abnormalities, coexisting systemic disease, previous ocular surgery, preterm birth, and difference in best-corrected visual acuity between the eyes greater than two lines. Informed consent was obtained from parents before participation. The research protocol was approved by the Institutional Review Board of the University of Texas Southwestern Medical Center and was performed according to the tenets of the Declaration of Helsinki. Measures of Binocular Vision Random Dot Stereopsis. Random dot stereoacuity was measured using the Randot Butterfly Test (Stereo Optical, Chicago, IL); the Randot (Version 2) Shapes Test (Stereo Optical); the Randot Preschool Stereoacuity Test
Fawcett and Birch
257
(Stereo Optical); and the Lang 1 Test (Fresnel Prism and Lens, Scottsdale, AZ). Coarse stereopsis was measured using the Titmus Fly Test (Stereo Optical). All tests were administered according to the protocol standardized by Pediatric Eye Disease Investigator Group (PEDIG) for multicenter clinical trials. Any children who were unable to cooperate with shape-based tests were assessed using the Infant Random Dot Stereoacuity Cards (Stereo Optical).9 A 2AFC preferential looking technique using a two down– one up staircase format beginning with 435 seconds of arc disparity was used to quantify stereoacuity when the Infant Random Dot Stereoacuity Cards were used. The session concluded when eight staircase reversals were obtained; when the child made at least six errors at the largest disparity; or when the first six trials were correct at the smallest disparity. Stereoacuity threshold was determined as the average disparity of the last six reversals.9 The normality of individual patient’s stereoacuity scores was determined using normative values derived using these tests.9,10 On the Randot Preschool Stereoacuity Test, 95% of children age ⱖ 3 years scored ⱖ 200 seconds of arc, and 95% of children age ⱖ 4 years scored ⱖ 100 seconds of arc.10 On the Infant Random Dot Stereoacuity Cards, 95% of children scored ⱖ 150 seconds of arc by 24 months of age.9 For this study, stereoacuity values greater than 95% tolerance limits for the test administered were considered to be abnormal. Motor Fusion/Fusional Vergence. Motor fusion/fusional vergence was measured using the 4 PD base-out test. A 4 PD loose prism was applied base out in front of one eye while the patient was instructed to fixate a small toy held approximately 33 cm out. Sensory foveal fusion was considered to be present in patients who produced a fusional vergence eye movement by the fellow eye in response to placement of the prism in front of the test eye. Grading for the 4 PD base-out test was pass/fail. To minimize the occurrence of false-negative results, passing required the ability to elicit a fusional vergence eye movement response in both eyes. Sensory Fusion. Sensory foveal fusion was measured using the Worth 4-dot test at 3 m (0.66°) according to the PEDIG standardized protocol. Grading of the Worth 4-dot test was pass/fail. Passing required reporting four dots. Failure required reporting two, three, or five dots. Only children who were able to cooperate or complete testing were included in the analysis of sensory foveal fusion (N ⫽ 50). Motion Visual Evoked Potential (mVEP). Nasotemporal mVEP asymmetry is associated with an absence of foveal fusion in patients with infantile ET and accommodative ET.11 In the current study, mVEP is included as an additional objective measure of binocularity. The NuDiva sweep-VEP system was used to acquire and analyze the mVEPs.12 Motion VEPs were elicited by jittering 1 cm/degree sine wave vertical gratings separated by 90° spatial phase at 6 Hz (12 reversals/sec). Gratings
258
Fawcett and Birch
were displayed on a 34° ⫻ 25° field using a high-resolution video monitor with a mean luminance of 161.6 candela/m2 and a contrast of 80%. Participants viewed the stimuli monocularly from a distance of 50 cm. Monocular viewing was achieved by covering one of the eyes with an opaque patch. During a single 10-second trial, mVEPs were recorded when the corneal reflection of the video monitor was centered in the child’s pupil. Recording was interrupted when fixation was lost. Fixation of the stimulus was maintained by dangling a small toy immediately in front of the monitor. The electroencephalogram (EEG) was recorded from two bipolar derivations (01 and 02), 2 to 2.5 cm to the left and right of a common reference electrode (0z) on the midline. A ground electrode was placed 2.5 cm above the reference electrode. The potential differences were amplified (gain ⫽ 10,000; ⫺3db cutoff at 1 and 100 Hz). The EEG was adaptively filtered (bandpass) in real time (sampling time ⫽ 397 Hz) to isolate the VEP. The EEG was subjected to Fourier analysis to extract the amplitude and phase of the VEP at the first (F1) and second (F2) harmonics of the stimulus. All analyses were based on vector averages of five or more trials for each eye. Amplitude, phase, signal-to-noise ratio, and a T 2 circ statistic were computed based on the vector average of the individual trials for each eye.13 The amplitude at the response frequency was compared with the average amplitude recorded at the noise frequencies adjacent to the recording frequency. A response was considered to be significant if its signal-to-noise ratio was ⬎ 3:1. Symmetric and asymmetric mVEPs produce characteristic Fourier spectra. A symmetric mVEP yields a Fourier spectra composed primarily of F2, where there is a peak of activity at twice the temporal frequency of the stimulus, ie, equal responses to the onset of nasal and temporalward horizontal motion. An asymmetric mVEP yields a Fourier spectra composed primarily of F1, where there is a peak of activity at the fundamental frequency of the stimulus, ie, a stronger response to the onset of motion in one direction of motion than the other. The presence of motion asymmetry was determined using the presence versus absence of a “bowtie.”11,12 By this method, motion asymmetry is distinguished by a significant F1 response component in the same channel(s) for both eyes with an interocular phase differences of 180 ⫾ 40°. Risk Factors Examined High AC/A Relationship. AC/A relationship was classified, using the heterophoria method, as normal or high according to the difference between the distance (6 m) and near (33 cm) prism and alternate cover measurements of eye alignment. Measures of distance and near eye alignment were measured using an accommodative target with hyperopic correction. A dichotomous classification of high AC/A was given for any patients with ⬎ 10 PD difference
Journal of AAPOS Volume 7 Number 4 August 2003
between the distance and near measurements of eye alignment and normal AC/A for any patients with ⱕ 10 PD difference between the distance and near measurements of eye alignment. Note that the heterophoria method, although commonly used in the clinic, does not distinguish between high AC/A ratio and nonaccommodative convergence excess.14 The range of difference between the distance and near measurements of eye alignment was between 0 and 45 PD. Twenty patients were classified with a high AC/A and 49 patients were classified with a normal AC/A ratio. High Hyperopia. Hyperopia was considered moderate to high if it was ⱖ 4.50 D spherical equivalent. Thirtyseven patients had hyperopia ⱖ 4.50 D, and 32 patients had hyperopia ⬍ 4.50 D. Anisometropia. Anisometropia was defined as ⱖ 1 D spherical equivalent refractive error difference between the two eyes. Fifteen patients had anisometropia, and 54 patients had no anisometropia. Age of Onset. Onset age was based on the parents’ and ophthalmologist’s reports. For the purpose of analysis, onset age was grouped as 0 to 6 months (congenital) versus ⱖ 7 months (acquired) and 0 to 17 months (infantile) versus ⱖ 18 months (late onset). Ten patients had onset at 0 to 6 months, and 21 patients had onset at 0 to 17 months. Duration of Eye Misalignment. Duration of eye misalignment was calculated as the difference between the onset age of constant eye misalignment and patient age at eye realignment. For the purpose of analysis, duration of constant eye misalignment was grouped as 0 to 3 months versus ⱖ 4 months and 0 to 11 months versus ⱖ 12 months. Twenty-three patients had a constant eye misalignment lasting 0 to 3 months, and 35 patients had a constant eye misalignment lasting 0 to 11 months. Statistics Risk Factor Analysis. The relationship between each risk factor and abnormal binocular vision outcome was calculated via the Mantel-Haenszel odds ratio (OR): OR ⫽ [P1/(1-P1)]/[P0/(1-P0)], where P1 is the incidence of the outcome in patients exposed to the risk factor, and P0 is the incidence of the outcome in patients not exposed to the risk factor. Because the incidence of abnormal binocular vision outcomes in this study population exceeded 10%, the Mantel-Haenszel odds ratios were corrected to relative risks (RR):15,16 RR ⫽ OR/(1-P0) ⫹ (P0 ⫻ OR) Relative risk data—along with 95% confidence intervals (CI)—are presented in Figures 1 and 3 through 6. Factors with a 95% CI lower limit ⬎ 1.0 pose a significant risk for an anomalous binocular vision outcome (equivalent to P ⬍ .05). Because the RR and its associated 95% CI cannot be computed when the observed incidence of an adverse out-
Journal of AAPOS Volume 7 Number 4 August 2003
come is equal to zero, for the purpose of analyses, any zeros were converted to 0.5 to provide a conservative estimate of the RR. Secondary Analysis. When a factor was found to pose a significant risk for abnormal stereoacuity, a secondary analysis was conducted to compare median stereoacuities of subgroups via the Kruskal-Wallis one-way analysis of variance (ANOVA) on ranks. When a factor was found to pose a significant risk for abnormal fusion (4 PD base-out test, Worth 4-dot test, or mVEP test), a secondary analysis was conducted to determine the significance of proportions of patients within subgroups either passing or failing.
RESULTS Stereoacuity scores ranged from 40 seconds of arc to nil. Binocular vision results are summarized in the Table 1. Because binocular vision is not an all-or-none sensory function, it is common to have different results on measures of sensory fusion, motor fusion, and stereopsis. It is believed that different tests measure different degrees of sensory and/or motor fusion and that different tests tap different underlying neural mechanisms of binocularity. The RR for each of the risk factors for abnormal stereopsis are summarized in Figure 1. The only clinical factor studied that posed a significant risk for abnormal stereopsis was duration of constant eye misalignment. Children with eye misalignment lasting ⱖ 4 months were 4.6 times more likely to have abnormal stereopsis than those with constant eye misalignment lasting 0 to 3 months. Kruskal-Wallis one-way ANOVA on ranks examining the difference between median stereoacuity scores by patients with 0 to 3 months versus 4 to 11 months versus ⱖ 12 months of constant eye misalignment showed a significant effect of duration on stereoacuity outcome (H ⫽ 54.373, P ⬍ .001). As shown in Figure 2, significant differences between the median stereoacuity scores were found between patients with 0 to 3 months versus 4 to 11 months and 0 to 3 months versus ⱖ 12 months of eye misalignment. The RRs of each of the risk factors for nil stereopsis are summarized in Figure 3. Duration of constant eye misalignment was the only clinical factor studied that posed a significant risk for nil stereopsis. Children with eye misalignment lasting ⱖ 4 months were 33 times more likely to have nil stereopsis than were children with 0 to 3 months of constant eye misalignment. There was a trend toward a higher proportion of patients with nil stereopsis in the group with 4 to 11 months of constant eye misalignment versus the group with 0 to 3 months of constant eye misalignment (25% versus 0%, z ⫽ 1.87, P ⫽ .06). Significant differences were found between the proportions of patients with nil stereopsis with 0 to 3 months versus ⱖ12 months of constant eye misalignment (0% versus 85%, z ⫽ 6.05, P ⬍ .001) and between 4 to 11 months versus ⱖ 12 months of constant eye misalignment (25% versus 85%, z ⫽ 3.54, P ⬍ .001).
Fawcett and Birch
259
TABLE 1. Incidence of abnormal binocular vision Random dot stereoacuity (secs arc) 40-60 100-250 9 16 (8 normal)*
400-3000 12
nil 32
Motor fusion/fusional vergence (4 PD base-out prism test) Pass Fail 32 37 Sensory fusion (Worth 4-dot test at 3 m) Pass Fail 27 23 Motion VEP Symmetric 49
Unable 19
Asymmetric 20
*Two hundred seconds of arc is within the normal range for 3-year-old children, and 100 seconds of arc is within the normal range for 4-year-old children; VEP, visual-evoked potential.
FIG 1. Relative risks for abnormal stereopsis as a function of high AC/A ratio (⬎10 D difference between near and distance eye alignment); hyperopia (ⱖ 3 D); anisometropia (ⱖ 1 D spherical equivalent refractive error difference between the two eyes); age at onset of constant ET (0 to 6, 0 to 17, and 0 to 24 months); and duration of constant ET (ⱖ 4 and ⱖ 12 months). The 95% CIs for relative risk are shown as horizontal bars. The relative risk for each of the clinical factors is shown as a vertical line within each horizontal bar. Significant risk is observed with 95% CI lower limit ⱖ 1.0.
The RR associated with each of the clinical factors for an absence of fusional vergence is shown in Figure 4. Both high AC/A ratio and duration of constant eye misalignment posed significant risks for failing the 4 PD base-out test. Children with a high AC/A ratio were 2.2 times more likely to fail the 4 PD base-out test than were children with a normal AC/A ratio. Children with eye misalignment lasting ⱖ 4 months were 37 times more likely to fail the 4 PD base-out test than were children with 0 to 3 months of constant eye misalignment. The difference between the proportions of patients who failed the 4 PD base-out test with 0 to 3 months versus 4 to 11 months of constant eye misalignment was statistically significant (0% versus 50%, z ⫽ 3.25, P ⬍ .001). Also significant were the differences in 4 PD base-out test failure (1) between the patients with 0 to 3 months versus ⱖ 12 months of constant eye misalignment (0% versus 88%, z ⫽ 6.26, P ⬍ .001) and (2)
260
Fawcett and Birch
Journal of AAPOS Volume 7 Number 4 August 2003
FIG 4. Relative risks for abnormal fusional vergence (failing the 4 PD base-out test) as a function of each of the risk factors. Other details as in Figure 1. FIG 2. Random dot stereoacuity (seconds of arc) as a function of duration of constant ET. Shaded areas indicate the interquartile range (25th and 75th percentile), and error bars indicate 10th and 90th percentiles for each duration group. The asterisks indicate a significant difference of median stereoacuity compared with the 0- to 3-month group as determined by Dunn’s pairwise comparisons (P ⬍ .05). The horizontal lines in the shaded boxes represent the median stereoacuity values (100 and 1775 seconds of arc). The median stereoacuity value for patients with ⱖ12 months of constant ET is nil.
FIG 5. Relative risks for absence of foveal fusion (failing the Worth 4-dot test at 3 months) as a function of each of the risk factors. Other details as in Figure 1.
FIG 3. Relative risks for nil stereopsis as a function of each of the risk factors. Other details as in Figure 1.
between 4 to 11 months versus ⱖ 12 months of constant eye misalignment (50% versus 88%, z ⫽ 2.33, P ⫽ .02). The RR associated with each of the clinical factors for an absence of sensory foveal fusion subsequent to successful eye realignment is shown in Figure 5. Duration of constant eye misalignment is the only clinical factor studied that posed a significant risk for failing the Worth 4-dot test at 3 m. Children with eye misalignment lasting ⱖ 4 months were 31 times more likely to fail the Worth 4-dot test at 3 m than were children with 0 to 3 months of constant eye misalignment. The difference between the proportions of patients who failed the Worth 4-dot test with 0 to 3 months versus 4 to 11 months of constant eye misalignment was statistically significant (0% versus 37.5%, z ⫽ 2.22, P ⬍ .03). Also significant were the differences in Worth 4-dot test failure (1) between the patients with 0 to 3 months versus ⱖ 12 months of con-
FIG 6. Relative risks for mVEP asymmetry as a function of each of the risk factors. Other details as in Figure 1.
stant eye misalignment (0% versus 91%, z ⫽ 5.59, P ⬍ .001) and (2) between 4 to 11 months versus ⱖ 12 months of constant eye misalignment (37.5% versus 91%, z ⫽ 2.58, P ⫽ .01). The RR associated with each of the clinical factors for an asymmetric mVEP subsequent to successful eye realignment is shown in Figure 6. Duration of constant eye misalignment was the only clinical factor studied that posed a significant risk for an asymmetric mVEP. Children with eye misalignment lasting ⱖ 4 months were 17 times more likely to have an asymmetric mVEP than children with 0 to 3 months of constant eye misalignment. The
Journal of AAPOS Volume 7 Number 4 August 2003
difference between the proportions of patients who exhibited an asymmetric mVEP with 0 to 3 months versus 4 to 11 months of constant eye misalignment was not statistically significant (0% versus 9%, z ⫽ 0.35, P ⫽ .73). The difference between the proportions of patients who exhibited an asymmetric mVEP with 0 to 3 months versus ⱖ 12 months of constant eye misalignment (0% versus 47%, z ⫽ 3.52, P ⬍ .001) was significant, and a similar trend was found when comparing patients with 4 to 11 months versus ⱖ 12 months of constant eye misalignment (9% versus 47%, z ⫽ 2.58, P ⫽ .055).
DISCUSSION In this study, we examined the RR of high AC/A relationship, high hyperopia, anisometropia, age at onset, and duration of constant eye misalignment on binocular vision outcomes in nonamblyopic children with accommodative ET, age 3 to 12 years, after successful treatment. Binocular vision was assessed using four measures: stereopsis, motor fusion/fusional vergence, sensory foveal fusion, and mVEP response. A high AC/A relationship was found to pose a significant risk only for an absence of fusional vergence as measured by the 4 PD base-out test. Within this sample of patients, a high AC/A relationship was not found to pose a significant risk for anomalous binocularity on other measures of binocularity including stereopsis, sensory foveal fusion, and the symmetry of the mVEP response. The lack of association between high AC/A relationship and abnormal stereopsis is in agreement with an earlier study.2 Both studies showed that patients with high AC/A ratios can achieve fine stereopsis with early bifocal spectacle and/or surgical intervention. Conversely, it is important to note that the cohort studied excluded any patients who had not achieved eye alignment within 8 PD at the time of testing. Had we included patients who had not been successfully aligned, the prevalence of nil stereopsis would have been much higher. Because the prevalence of poor alignment may be higher among children with high AC/A ratios,3-6 we may have underestimated the overall risk of high AC/A relationship on poor binocular sensory outcomes in the general clinical population. High hyperopia was not found to pose a significant risk for abnormal binocular vision among patients with accommodative ET. In our sample of patients, hyperopia ranged between 0.25 and 8 D, with an equal proportion of patients falling within the categories of mild hyperopia (44%) and moderate to high hyperopia (56%). Within these categories, there was an equal distribution of children with abnormal versus normal outcomes on each of the measures of binocular vision. Anisometropia was not found to pose a significant risk for abnormal binocular vision in patients with accommodative ET. The small representation of patients with anisometropia within this sample (22%), however, decreases our confidence in making any strong conclusions about
Fawcett and Birch
261
this finding. A larger sample of patients with accommodative ET who have anisometropia is necessary to make a reliable calculation of the RR of anisometropia for abnormal binocular vision. Age at onset was not found to pose a significant risk for abnormal binocular vision for patients with accommodative ET whose eyes were successfully realigned by the time of testing. Infantile-onset versus childhood-onset accommodative ET was found to pose no greater risk for an abnormal binocular vision outcome after successful eye realignment. A delay of surgical alignment of infantile ET longer than 3 months results in poor stereoacuity outcomes.17 A similar observation has been made in a cohort of children with accommodative ET.2 Here we show that other measures of binocular sensory function (fusional vergence, sensory foveal fusion, and mVEP response) share a similar duration of abnormal eye alignment for disruption. Prolonged eye misalignment can be the result of a number of factors including delayed medical referrals from primary health care providers, third-party payer requirements for second opinions, poor compliance with treatment, and an initially ineffective course of treatment. An important consideration in treatment of strabismus is understanding the importance of prompt eye alignment to rescue binocular visual function. To expand our understanding of the time course of the critical period of development, continued susceptibility, and recovery of binocular vision, we examined the differences of binocular sensory function in children with accommodative ET who experienced 0 to 3 months of eye misalignment, 4 to 11 months of eye misalignment, and longer than 12 months of eye misalignment before being successfully treated. Significant differences between the stereoacuity scores of successfully realigned patients with 0 to 3 months (100 seconds of arc) versus 4 to 11 months (1775 seconds of arc) of constant ET indicate that with increasing duration of eye misalignment, the amount of stereoacuity that can be achieved after successful treatment decreases. An analysis of the differences between the proportions of patients with no measurable stereopsis who have 0 to 3 months (0%) versus 4 to 11 months (25%) versus ⱖ 12 months (85%) of constant eye misalignment indicates that if not moderate stereopsis, at least coarse stereopsis can be achieved when treatment is effective in correcting eye alignment after up to 11 months of constant eye misalignment. Fusional vergence, sensory foveal fusion, and symmetric mVEP responses are equally, if not more susceptible to loss, as is stereopsis. An analysis of the differences between the proportions of patients with no fusional vergence, no sensory foveal fusion, and asymmetric mVEPs, respectively, who have 0 to 3 months (0%, 0%, and 0%) versus 4 to 11 months (50%, 37.5%, and 9%) versus ⱖ 12 months (88%, 91%, and 47%) of constant eye misalignment indicates that with increasing duration of eye misalignment, fusional vergence, sensory foveal fusion, and symmetric mVEP responses are lost.
262
Fawcett and Birch
Overall, the data presented here support the hypothesis that the duration of constant eye misalignment is the primary risk factor for adverse binocular vision outcomes after successful treatment of children with accommodative ET. References 1. Parks MM. Ocular motility and strabismus. Hagerstown (MD): Harper and Row; 1975. p. 99-105. 2. Fawcett SF, Leffler JN, Birch EE. Factors influencing stereoacuity in accommodative esotropia. J AAPOS 2000;4:15-20. 3. Parks MM. Management of acquired esotropia. Br J Ophthalmol 1974;58:240-7. 4. Pratt-Johnson JA, Tilson G. Sensory outcome with nonsurgical management of esotropia with convergence excess (a high accommodative convergence/accommodation ratio). Can J Ophthalmol 1984;19: 220-3. 5. Ludwig IH, Parks MM, Getson PR, Kammerman LA. Rate of deterioration in accommodative esotropia correlated to the AC/A relationship. J Pediatr Ophthalmol Strabismus 1988;25:8-12. 6. Wilson ME, Bluestein EC, Parks MM. Binocularity in accommodative esotropia. J Pediatr Ophthalmol Strabismus 1993;30:233-6. 7. Weakley DR, Birch EE. The role of anisometropia in the development of accommodative esotropia. Trans Am Ophthalmol Soc 2000; 98:71-9.
Journal of AAPOS Volume 7 Number 4 August 2003 8. Weakley DR, Birch EE, Kip K. The role of anisometropia in the development of accommodative esotropia. J AAPOS 2001;5:153-7. 9. Birch EE, Solomao SR. Infant random dot stereocards. J Pediatr Ophthalmol Strabismus 1998;35:86-90. 10. Birch EE, Williams C, Hurtes J, Lapa MC, ALSPAC. Children in Focus study team. Random dot stereoacuity of preschool children. J Pediatr Ophthalmol Strabismus 1997;34:217-22. 11. Fawcett SF, Birch EE. Motion VEPs, stereopsis, and bifoveal fusion in children with strabismus. Invest Ophthalmol Vis Sci 2000;41: 411-6. 12. Norcia AM, Garcia H, Humphrey R, Holmes A, Hamer RD, OrelBixler D. Anomalous motion VEPs in infants and in infantile esotropia. Invest Ophthalmol Vis Sci 1991;32:436-9. 13. Victor JD, Mast J. A new statistic for steady state evoked potentials. Electroencephalogr Clin Neurophysiol 1991;78:378-88. 14. von Noorden GK, Campos EC. Binocular vision and ocular motility. 6th ed. St. Louis: Mosby; 2002. p. 85-100. 15. Zang J, Yu KF. What’s the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA 1998;280: 1690-1. 16. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22: 719-48. 17. Birch EE, Fawcett S, Stager DR. Why does early surgical alignment improve stereoacuity outcomes in infantile esotropia? J AAPOS 2000;4:10-14.