Optical Treatment of Amblyopia in Astigmatic Children

Transcription

1 Optical Treatment of Amblyopia in Astigmatic Children The Sensitive Period for Successful Treatment Erin M. Harvey, PhD, 1,2 Velma Dobson, PhD, 1,3 Candice E. Clifford-Donaldson, MPH, 1 Joseph M. Miller, MD, MPH 1,2,4 Objective: To compare the effectiveness of eyeglass treatment of astigmatism-related amblyopia in children younger than 8 years (range, years) versus children 8 years of age and older (range, years) over short (6-week) and long (1-year) treatment intervals. Design: Prospective, interventional, comparative case control study. Participants: Four hundred forty-six nonastigmatic (right and left eye, 0.75 diopters [D]) and 310 astigmatic (RE, 1.00 D) Native American (Tohono O odham) children in kindergarten or grades 1 through 6. Intervention: Eyeglass correction of refractive error, prescribed for full-time wear, in astigmatic children. Main Outcome Measures: Amount of change in mean right-eye best-corrected letter visual acuity for treated astigmatic children versus untreated, age-matched nonastigmatic children after short (6-week) and long (1-year) treatment intervals. Results: Astigmatic children had significantly reduced mean best-corrected visual acuity at baseline compared to nonastigmatic children. Astigmats showed significantly greater improvement in mean best-corrected visual acuity (0.08 logarithm of the minimum angle of resolution [logmar] unit; approximately 1 line), than the nonastigmatic children (0.01 logmar unit) over the 6-week treatment interval. No additional treatment effect was observed between 6 weeks and 1 year. Treatment effectiveness was not dependent on age group ( 8 years vs. 8 years) and was not influenced by previous eyeglass treatment. Despite significant improvement, mean best-corrected visual acuity in astigmatic children remained significantly poorer than in nonastigmatic children after 1 year of eyeglass treatment, even when analyses were limited to results from highly compliant children. Conclusions: Sustained eyeglass correction results in significant improvement in best-corrected visual acuity in astigmatic children, including those previously believed to be beyond the sensitive period for successful treatment. Ophthalmology 2007;114: by the American Academy of Ophthalmology. Disruption of normal visual experience during early development can lead to amblyopia, neural visual deficits characterized by reduced vision that persists after normal visual input is restored. Uncorrected astigmatism is one condition that can cause abnormal visual experience during early development and can result in amblyopia. Several studies have documented reduced best-corrected visual acuity associated with astigmatism (Invest Ophthalmol Vis Sci 25[Suppl]:217, 1984; Dissertation Abstracts International 63: No 12B p6115, 2002; Invest Ophthalmol Vis Sci 45:Eabstract 2580, 2004). 1 4 In addition, presence and severity of visual deficits in astigmatism-related amblyopia can be specific to stimulus orientation, 1,3 13 a specific form of amblyopia termed meridional amblyopia, the result of the orientation-specific defocus characteristic of uncorrected astigmatism. The most common treatment of astigmatism-related amblyopia that occurs in the absence of strabismus or anisometropia is restoration of normal visual input through optical correction of refractive error. Few studies have focused specifically on the effectiveness of optical treatment of astigmatism-related amblyopia, but the available literature Originally received: September 19, Final revision: March 7, Accepted: March 8, Manuscript no Department of Ophthalmology and Vision Science, University of Arizona, Tucson, Arizona. 2 College of Public Health, University of Arizona, Tucson, Arizona. 3 Department of Psychology, University of Arizona, Tucson, Arizona. 4 College of Optical Sciences, University of Arizona, Tucson, Arizona. Presented in part at: Optical Society of America Fall Vision Meeting, October 2005, Tucson, Arizona. Supported by the National Eye Institute, Bethesda, Maryland (grant nos. EY11155 [JMM], EY13153 [EMH]), and Research to Prevent Blindness, Inc., New York, New York (unrestricted grant to the University of Arizona Department of Ophthalmology and Vision Science [JMM], Walter E. and Lilly Disney Award for Amblyopia Research [JMM], and Career Development Award [EMH]). There are no conflicting relationships between any author and any product mentioned. Correspondence to Erin M. Harvey, PhD, University of Arizona Department of Ophthalmology and Vision Science, 655 North Alvernon Way, Suite 108, Tucson, AZ emharvey@u.arizona.edu by the American Academy of Ophthalmology ISSN /07/$ see front matter Published by Elsevier Inc. doi: /j.ophtha

2 Ophthalmology Volume 114, Number 12, December 2007 suggests that successful treatment is limited to a sensitive period. Retrospective studies, based on a small number of astigmats, have found that astigmatic adults do not show evidence of meridional amblyopia if astigmatism is corrected with eyeglasses before the age of 7 years and are more likely to show evidence of meridional amblyopia if the astigmatism is corrected after age 7 years. 6,12 Consistent with these findings, Mohindra et al 7 reported that a 3-yearold astigmatic child with meridional amblyopia on initial testing showed no evidence of meridional amblyopia after 3 months of eyeglass wear. However, in a prospective study of eyeglass treatment of astigmatism-related amblyopia, 14 we found that astigmatic 3- to 5-year-old Native American children had significantly poorer mean best-corrected letter visual acuity than nonastigmatic children, and there was no significant improvement in best-corrected visual acuity after an average eyeglass treatment duration of 4 months. These results suggest that even when astigmatism is corrected as early as age 3 years, astigmatism-related amblyopia may persist. However, several alternative explanations of the null treatment effect, including high measurement variability (because the children were cyclopleged during acuity testing and were wearing trial frames), variability in treatment compliance, and insufficient treatment duration, prevented a definitive conclusion that these children were beyond the sensitive period for successful treatment of astigmatism-related amblyopia. The goal of the present study was to define more clearly the sensitive period for successful treatment of astigmatismrelated amblyopia. We compared effectiveness of eyeglass treatment in 2 cohorts of astigmatic children: a younger cohort (YC; younger than age 8 years; age range, almost 5 to 8 years) and an older cohort (OC; age range, 8 to 14 years). If the sensitive period for successful treatment ends before 3 to 5 years of age, 14 a treatment effect in either cohort should not be demonstrated. If the sensitive period extends only to approximately age 7 years, 6,12 a treatment effect only in the YC should be evident. Finally, significant treatment effects in both cohorts would suggest that the sensitive period extends beyond age 7 years. This finding would be contrary to previous small-sample reports regarding treatment of astigmatism-related amblyopia. 6,12,14 Patients and Methods Patients The present study focused on members of a Native American tribe, the Tohono O odham, because previous research has shown that there is a high prevalence of astigmatism and astigmatismrelated amblyopia 2,3,14 among members of the tribe. Subjects eligible for recruitment and participation were all children in kindergarten through grade 2 (recruited during the school year) and all children in grades 4 through 6 (recruited during the school year) who attended 1 of 5 elementary schools on the Tohono O odham Reservation in southern Arizona. In addition, data were included from children at a sixth elementary school on the reservation who participated in a preliminary study during the and school years in which an identical inclusion, treatment, and outcome assessment protocol was used (Dissertation Abstracts International 63: No 12B p6115, 2002). Recruitment dates for the 2 cohorts were designed to minimize the possibility of recruiting children who participated in a previous eyeglass treatment study of preschool children from the same population. 3,18,19 Children who did not enroll during the first year of testing of their cohort but were in targeted grades (kindergarten grade 2 or grades 4 6) during the second year of testing for their cohort were invited to participate and contributed baseline and initial follow-up data, but 1-year follow-up testing was not conducted during the following year. Although all children in the targeted grades were invited and were eligible to participate in the study, only data from subjects meeting specific refractive error and ophthalmic criteria were included in analyses (see Data Analysis and Predictions ). The research was approved by the Institutional Review Board of the University of Arizona and was compliant with the Health Insurance Portability and Accountability Act. Written informed consent was obtained from a parent or guardian prior to a child s participation, and written assent was obtained from children in grades 4, 5, and 6. Procedures A flow chart of study procedures is provided in Figure 1. Each child underwent a complete baseline eye examination, including cycloplegic refraction, conducted by a pediatric ophthalmologist (JMM) at least 40 minutes after instillation of 1 drop of proparacaine (0.5%) and 2 drops of cyclopentolate (1%) separated by an interval of approximately 5 minutes. Eyeglasses were prescribed for children who had 2.00 diopters (D) or more of astigmatism in either eye, and children who had uncorrected letter acuity worse than 20/20 and met 1 or more of the following criteria: myopia of 0.75 D or more in either meridian, hyperopia of 2.50 D or more in either meridian, astigmatism of 1.00 D or more in either eye, anisometropia of 1.50 D or more spherical equivalent. Refractive error was determined by cycloplegic autorefraction (Nikon Retinomax K ; Nikon Inc., Tokyo, Japan; now manufactured by Righton Manufacturing Co., Tokyo, Japan), confirmed by cycloplegic retinoscopy and, when possible, subjective refinement. Approximately 2 to 3 weeks after the baseline eye examination, each child participated in baseline best-corrected letter visual acuity testing. Although eyeglasses were prescribed only for children who met the above guidelines, all children wore eyeglasses containing their refractive correction (based on cycloplegic autorefractor measurements, refined or confirmed by cycloplegic retinoscopy or subjective refinement) for acuity testing. This allowed assessment of best-corrected visual acuity in all children and masked the testers as to which children had been prescribed eyeglasses. Children who did not meet the prescribing criteria wore a pair of eyeglasses from a set of stock eyeglasses that were selected with the restriction that both the right and left lens corrections were less than a 0.50 vector dioptric difference from the child s refractive error. 20,21 For both prescriptions and selection of stock eyeglasses, correction of hyperopic refractive error was reduced by one third or by 1.00 D, whichever was greater. 22 Full correction of astigmatism at the appropriate axis was given. At the beginning of the baseline acuity testing session, each child was fit with the appropriate set of eyeglasses. Acuity testing then was conducted by testers who were masked to each child s refractive error and to results obtained at previous testing sessions. Monocular best-corrected right eye (RE) and left eye distance logarithm of the minimum angle of resolution (logmar) letter acuity was determined using cm Early Treatment Diabetic Retinopathy Study charts 23 mounted in an illuminator cabinet (Precision Vision, Inc., LaSalle, IL) at 4 m. The Early Treatment Diabetic Retinopathy Study chart 1 (catalog item no. 2121) was used for testing the RE, which was always tested first, and chart

3 Harvey et al Treatment of Astigmatism-Related Amblyopia Figure 1. Flow chart showing session types during year 01 and year 02 testing of cohorts. During year 01, kindergarten, first-, second-, fourth-, fifth-, and sixth-grade children who enrolled participated in 3 sessions: baseline eye examination (refractive error determined by cycloplegic autorefraction confirmed or refined by retinoscopy), baseline vision testing (best-corrected letter acuity, tested in eyeglasses containing correction determined at baseline eye examination), and 6-week follow-up vision testing (best-corrected letter acuity, tested in eyeglasses containing correction determined at baseline eye examination). During year 02, children who enrolled in year 01 participated in two sessions: follow-up eye examination (refractive error determined by cycloplegic autorefraction confirmed or refined by retinoscopy), and 1-year follow-up vision testing (best-corrected letter acuity, tested in eyeglasses containing correction determined at follow-up eye examination). In addition, during year 02, additional children in the targeted grade range were allowed to enroll. These children completed baseline eye examination, baseline vision testing, and 6-week follow-up vision testing. (catalog item no. 2122) was used for testing the left eye. The child was asked to identify verbally or by matching to letters on a lap card all 5 letters on each line of the chart, until he or she could no longer identify any of the 5 letters on a line. Best-corrected visual acuity was scored as the smallest letter size on which the child correctly identified at least 3 of a maximum of 5 letters. Logarithm of the minimum angle of resolution visual acuity scores were used for data analyses. At the end of the baseline best-corrected visual acuity testing session, children who were prescribed eyeglasses were given their eyeglasses to wear and to take home and stock glasses were collected from children who were not prescribed eyeglasses. A follow-up acuity testing session was conducted approximately 6 weeks after the baseline session (targeted follow-up interval was 1 month after baseline testing, or shortly thereafter), and a second follow-up acuity testing session was conducted approximately 1 year after the baseline acuity testing session. Refractive error correction for the baseline and 6-week follow-up vision testing sessions was based on the initial eye examination results. Refractive error correction for the 1-year follow-up vision testing session was based on the 1-year follow-up eye examination (identical to the baseline examination) conducted approximately 2 to 3 weeks before the 1-year follow-up visual acuity testing session. Encouraging and Monitoring of Treatment Compliance On the day eyeglasses were initially dispensed (the baseline acuity testing session), children who required eyeglasses were given a pair and were instructed to wear them all the time. A spare pair of eyeglasses was given to the child s teacher to keep in the classroom. Teachers were asked to give the spare eyeglasses to the child if they forgot to bring their other pair to school and to try to collect 2295

4 Ophthalmology Volume 114, Number 12, December 2007 them at the end of the day so that the child would always have a pair at school. Eyeglasses were repaired or replaced whenever they became lost, broken, or badly scratched. A study staff member responsible for monitoring eyeglass wear carried a spare for each child so that they could be dispensed as soon as needed. As soon as this spare was dispensed, a replacement spare was immediately ordered. At the end of the school year, the child was given the classroom spare to take home and was instructed to use them if their other pair became lost or broken over the summer break. Eyeglass wear was monitored by a study staff member who made periodic visits to classrooms and by teachers who were given daily log books to record whether children were wearing their eyeglasses. The log books also served as a reminder to teachers to dispense the classroom pair if the child did not have his or her eyeglasses. Data Analysis and Predictions All visual acuity analyses were based on RE data. Subjects were assigned to groups based on the cycloplegic refraction results of the baseline eye examination. The nonastigmatic control group was composed of children with little or no astigmatism ( 0.75 D in the RE and the left eye), and the astigmatic group was composed of children with high RE with-the-rule astigmatism ( 1.00 D; axis, ) and uncorrected letter acuity worse than 20/20 (children with 20/20 or better uncorrected visual acuity were not prescribed eyeglasses [i.e., were untreated] per study protocol). Data from subjects who did not meet the criteria for the nonastigmatic or the astigmatic group, subjects with any ocular abnormalities (other than high refractive error), and subjects with anisometropia ( 1.50 D interocular difference in spherical equivalent) were excluded from analyses. The astigmatic group was divided further into previously untreated children (no history of eyeglass wear, per parent report) and previously treated children (positive history of eyeglass wear, per parent report). Subjects also were categorized by age group: children younger than 8 years (YC) versus children 8 years of age and older (OC). Primary analyses compared the mean change in best-corrected visual acuity over time across groups using a repeated-measures analysis of variance. Significant treatment effects were defined as significantly greater improvement in mean best-corrected visual acuity over time in the astigmatic group relative to the nonastigmatic group. A repeated-measures analysis of variance compared mean change in measurements of best-corrected visual acuity across cohorts (YC vs. OC) from baseline to 6 weeks to 1 year. To increase statistical power, particularly with regard to detecting subtle effects of previous treatment, cohort, or both on treatment effectiveness, separate analyses evaluating change from baseline to 6 weeks also were conducted because the sample size for this follow-up period was substantially greater than the sample size for the 1-year follow-up period. The difference in samples size primarily was because additional children were allowed to enter the study during the second year (when 1-year data were being collected on subjects enrolled during the first year), and these additional children provided baseline and 6-week data, but not 1-year data. Secondary analyses of best-corrected visual acuity data evaluated the influence of treatment compliance on outcome of treatment. All compliance data for each child were reviewed by 4 raters (study staff members). Raters provided a compliance rating on a scale from 1 to 5, with 1 representing never wore eyeglasses and 5 representing always wore eyeglasses. Because the amount of compliance data available varied across subjects (e.g., some teachers recorded in the log books consistently and others recorded in them rarely or not at all), the raters also provided a confidence rating for their compliance rating. Confidence was rated on a scale from 1 to 5, with 1 representing very low confidence or little data available and 5 representing very high confidence or a lot of data available. Children with mean compliance and mean compliance confidence ratings of 3 or more were considered highly compliant. Finally, secondary analysis of treatment outcome data was conducted to determine what proportion of children identified as amblyopes at baseline were responsive to treatment. Amblyopia was defined as RE best-corrected letter visual acuity of 20/40 or worse. For the astigmatic children identified as amblyopes at baseline, analyses focused on comparing the percentage of YC versus OC children who no longer met the criterion for amblyopia after 1 year of treatment. Results A total of 1048 kindergarten through second-grade and fourththrough sixth-grade children were enrolled in the study. Of these children, 250 were excluded from analyses for the following reasons: anisometropia and ocular abnormality (n 1), strabismus and ocular abnormality (n 1), history of patching (n 1), anisometropia and strabismus (n 2), undilated at initial examination (n 2; 1 declined, 1 dilated poorly), met the criteria for astigmatic group but no data on previous treatment were available (n 7), strabismus (n 11), ocular abnormality other than strabismus (n 11), anisometropia (n 18), lost to follow-up after baseline eye examination (no baseline best-corrected visual acuity data collected, n 39), or did not meet the criteria for either group (nonastigmatic or astigmatic, n 157). Of the 798 children who provided baseline best-corrected visual acuity data and met the inclusion criteria, 94.7% (756/798) were tested at 6 weeks (acuity testing was not completed on 4 children [ 1%], 38 [5%] were lost to follow-up). Both 6-week and 1-year follow-up data were collected for 67.5% of children (539/ 798) who provided baseline best-corrected visual acuity data (5 [1%] did not complete acuity testing at 1 year; 112 [14%] were lost to follow-up at 6 weeks, 1 year, or both; and 142 [18%] entered the study during the second year of testing for their cohort and contributed baseline and 6-week data, but testing at 1 year was not attempted because testing of their cohort had ended). For the sample of 756 children followed up for at least 6 weeks, 372 children were younger than 8 years of age and were assigned to the YC (mean age, 6.37 years; standard deviation (SD), 0.84 year; range, years), and 384 children were at least 8 years of age and were assigned to the OC (mean age, years; SD, 1.36 years; range, years). Of the 756 children who met the inclusion criteria and were tested at 6 weeks, 446 (59.0%) met the criteria for the nonastigmatic group and 310 (41.0%) met the criteria for the astigmatic group. Among astigmats, 93 (30.0%) had no previous eyeglass correction and 217 (70.0%) reported a positive history of eyeglass correction. However, only 58 (18.7%) astigmats were wearing eyeglass correction on initial examination. For the smaller sample of 539 children who were tested at both 6 weeks and 1 year, 320 (59.4%) met the criteria for the nonastigmatic group and 219 (40.6%) met the criteria for the astigmatic group (52 [23.7%] with no previous eyeglass correction and 167 [76.3%] with a positive history of eyeglass correction). The first follow-up vision testing session (6-week follow-up) occurred at an average of 45 days after the baseline vision testing session (range, days for the 6-week follow-up sample [n 756] and ranging from 18 to 96 days for the smaller sample of children followed 1 year [n 539]). The 1-year follow-up vision testing session was conducted an average of 376 days (range, days) after the baseline vision testing session. A summary 2296

5 Table 1. Sample Sizes and Mean Weeks from Baseline to Follow-up for All Children Followed up at the First (Short-Duration) Follow-up (n 756) and for Children Who Were Tested at Both the Short- and Long-Duration Follow-ups (n 539) Cohort Harvey et al Treatment of Astigmatism-Related Amblyopia Group N Short Follow-up Sample Weeks from Baseline to First Follow-up Mean Short and Long Follow-up Sample Weeks from Baseline to First Follow-up Standard Deviation N Mean Standard Deviation Weeks from Baseline to Second Follow-up Mean Standard Deviation Younger cohort ( 8 yrs) Nonastigmats* Astigmats Total Older cohort ( 8 yrs) Nonastigmats* Astigmats Total Total Nonastigmats* Astigmats Total *Right eye and left eye, 0.75 diopters of astigmatism. Right eye, 1.00 diopters of astigmatism. of mean length of follow-up interval for each group and for each cohort at 6 weeks and 1 year is provided in Table 1. Preliminary analyses indicated that there were significant differences among groups on length of follow-up. The mean 1-year follow-up interval did not differ by cohort or by astigmatism group, but for both the larger 6-week follow-up sample and the smaller 1 year follow-up sample, the mean baseline to 6-week interval was significantly longer for the YC than for the OC and was longer for astigmats than for nonastigmats (all P 0.01). Therefore, repeated-measures analyses of covariance (rather than repeated-measures analyses of variance) were conducted to include duration of follow-up as covariates. Baseline Best-Corrected Letter Visual Acuity An analysis of variance of baseline best-corrected visual acuity data yielded a significant effect across groups (F 2, , P 0.001) and a significant effect of cohort (F 1, , P 0.001), reflecting significantly better best-corrected visual acuity in the OC than in the YC. The interaction between group and cohort was not significant. Post hoc analyses comparing bestcorrected visual acuity across groups (with Bonferroni correction) indicated that best-corrected visual acuity was significantly poorer in the astigmatic groups (previously treated [mean logmar visual acuity, 0.24; SD, 0.17] and previously untreated [0.30; SD, 0.16; P 0.001]) in comparison with the nonastigmatic group (mean logmar visual acuity, 0.04; SD 0.15), and the previously untreated astigmats had significantly poorer mean best-corrected visual acuity than the previously treated astigmats (P 0.003). When analyses were restricted only to subjects who completed both the 6-week and 1-year follow-ups, the main effects of group and cohort remained significant, and the group by cohort interaction neared significance (P 0.06; Fig 2A). Post hoc analyses indicated that mean best-corrected visual acuity in treated astigmats was better than in untreated astigmats, but the difference no longer reached statistical significance. Change in Best-Corrected Letter Acuity: Baseline versus 6 Weeks Repeated-measures analyses of covariance yielded significant main effects of time (F 1, , P 0.007, reflecting significantly better best-corrected visual acuity at follow-up), group (F 2, , P 0.001, reflecting significantly better bestcorrected visual acuity in the nonastigmatic group than in the astigmatic groups [P 0.001] and better best-corrected visual acuity in the previously treated astigmats than in the previously untreated astigmats [P 0.003]), and cohort (F 1, , P 0.001, reflecting better best-corrected visual acuity in the OC than in the YC). There was a significant interaction between time and group (F 2, , P 0.001). Post hoc analyses indicated that there was significantly more improvement in the astigmatic groups (previously treated and previously untreated) than in the nonastigmatic group (P 0.001), but no difference between previously treated and previously untreated astigmats. The interaction between time, group, and cohort was not significant (P 0.41), indicating that there was no evidence of differences in treatment effect between the YC and the OC. Change in Best-Corrected Letter Acuity: Baseline versus 6 Weeks versus 1 Year (Fig 2A) Repeated-measures analyses of covariance on data from subjects who completed best-corrected letter acuity measurements at baseline, 6 weeks, and 1 year (n 539) yielded significant main effects of time (F 2, , P 0.05, reflecting significant improvement from baseline to 6 weeks and from 6 weeks to 1 year), group (F 2, , P 0.001, reflecting significantly better bestcorrected visual acuity in the non-astigmatic group than in the astigmatic groups), and cohort (F 1, , P 0.001, reflecting better best-corrected visual acuity in the OC than in the YC). There was a significant interaction between time and group (F 4, , P 0.001). Post-hoc analyses indicated that there was significantly greater mean improvement from baseline to 6 weeks in the astigmatic groups than in the non-astigmatic group (P 0.001), but no difference between previously treated and previously untreated astigmats. Amount of improvement from 6 weeks to 1 year did not differ significantly among groups. The interaction between time, group, and cohort was not significant (P 0.21), indicating that there was no evidence of differences in treatment effect between the YC and the OC. Because of the large range in the follow-up intervals for 6 weeks and 1 year (Table 1), a post hoc analysis was conducted to 2297

6 Ophthalmology Volume 114, Number 12, December 2007 and previously untreated astigmats. The amount of improvement from 6 weeks to 1 year did not differ significantly among groups. Final Best-Corrected Letter Visual Acuity (1 Year) An analysis of covariance comparing final best-corrected visual acuity (1 year after baseline testing) by group and cohort yielded main effects of group (F 2, , P 0.001), indicating that nonastigmats still had significantly better best-corrected visual acuity than astigmats (previously treated and previously untreated) after 1 year of treatment (P 0.001), and a main effect of cohort (F 1, , P 0.001), indicating that the YC still had poorer best-corrected visual acuity than the OC. The interaction between group and cohort was not significant (P 0.53). We examined final best-corrected visual acuity data for the nonastigmatic group in comparison with the astigmatic subjects who were rated as highly compliant with treatment (mean compliance rating, 3, and mean confidence in compliance rating, 3). As shown in Figure 2B, the results were similar to the overall sample. An analysis of variance of final best-corrected letter acuity indicated a significant main effect of group (F 2, ; P 0.001), indicating that even with analysis limited to compliant subjects, nonastigmats still had significantly better best-corrected visual acuity than astigmats (previously treated and previously untreated; P 0.001). Figure 2. Graphs showing the mean best-corrected letter acuity for children tested at baseline, 6 weeks, and 1 year for (A) the entire longitudinal (1-year) sample of nonastigmats and astigmats and (B) the entire longitudinal (1-year) sample of nonastigmats and compliant astigmats. Open symbols represent children in the younger cohort (YC); filled symbols represent children in the older cohort (OC). Means are presented for nonastigmats, astigmats with no history of previous treatment, and astigmats with a positive history of previous treatment. Bars indicate 1 standard error of the mean. determine if the pattern of results differed when analyses were restricted to include only subjects within tighter follow-up windows who are more representative of 6-week and 1-year treatment intervals. A repeated-measures analysis of covariance on data from subjects who met the following criteria was conducted: (1) completed best-corrected letter acuity measurements at baseline, 6 weeks, and 1 year; (2) had 6-week follow-up data collected from 4 weeks or more to 8 weeks or less after baseline; and (3) and 1-year follow-up data collected from 10 to 14 months ( 40 to 56 weeks) after baseline (n 341). The interaction between time and group remained significant (F 4, , P 0.003). Post hoc analyses indicated that there was still a significantly greater mean improvement from baseline to 6 weeks in the astigmatic groups than in the nonastigmatic group (P 0.03 for previously untreated astigmats; P for previously treated astigmats) and no difference in the amount of improvement between previously treated Clinical Response to Treatment (Table 2) For the YC, a significantly larger proportion of children in the astigmatic groups (previously treated and previously untreated) than in the nonastigmatic group met the criterion for amblyopia at baseline (best-corrected visual acuity of 20/40 or worse; P 0.001). For the OC, a significantly larger proportion of children in the previously treated astigmatic group than in the nonastigmatic group met the criterion for amblyopia (P 0.001). Although a larger percentage of OC children in the previously untreated astigmatic group than in the nonastigmatic group met the criterion for amblyopia at baseline, the difference did not reach significance. The percentage of amblyopes in the previously treated versus the previously untreated astigmatic groups at baseline did not significantly differ for either the YC or the OC. Among astigmatic children who met the criterion for amblyopia at baseline, there was a significant reduction in the proportion with amblyopia at 1 year for astigmatic children in the YC (both previously treated and previously untreated) and for previously treated astigmats in the OC (P 0.001). The reduction in amblyopia in the previously untreated OC astigmats was 100%, but did not reach statistical significance because of the small sample of previously untreated OC children who met the criterion for amblyopia at baseline (n 2). There was no evidence of a significantly reduced treatment effect in the OC as compared with the YC, nor was there a suggestion of the predicted cohort effect (a greater reduction in the proportion of amblyopes in the YC than in the OC). In fact, the percentage of astigmatic children who still met the criterion for amblyopia was greater in the YC than in the OC, for both previously untreated and previously treated astigmats, although this trend was not statistically significant. Discussion The results of the present study indicate that: (1) there was significantly greater improvement in best-corrected letter 2298

7 Harvey et al Treatment of Astigmatism-Related Amblyopia Table 2. Percentage of Children Who Met the Criterion for Amblyopia (Best-Corrected Visual Acuity of 20/40 or Worse) by Group and by Cohort at Baseline and at 1 Year for the Overall Longitudinal (1-Year) Sample and the Percentage of Children Who Met the Criterion for Amblyopia at Baseline Who Remained Amblyopic at 1 Year Baseline 1 Year Baseline Amblyopes Who Remained Amblyopic at 1 Year Cohort Group N n % n % N n % Younger cohort ( 8 yrs) Nonastigmats Previously untreated astigmats Previously treated astigmats Total Older cohort ( 8 yrs) Nonastigmats Previously untreated astigmats Previously treated astigmats Total N total no. of children in the group or cohort; n no. of amblyopes; % of amblyopes in the group or cohort (n/n 100). visual acuity from baseline to 6 weeks in the astigmatic groups than in the nonastigmatic group, indicating a significant effect of eyeglass treatment; (2) there were no significant treatment effects observed from 6 weeks to 1 year, suggesting that the observed treatment effects occurred primarily very early in treatment (within the first few weeks); (3) treatment effects did not differ between cohorts, that is, there was no evidence that older children were less responsive than younger children to treatment; and (4) astigmats who had been treated previously with eyeglass correction had better baseline best-corrected visual acuity than did astigmats who had not been previously treated, but previous treatment did not seem to reduce the improvement in bestcorrected visual acuity that occurred after eyeglass correction in the present study (treatment effect did not differ between previously treated and untreated astigmats). Analyses that examined reduction in prevalence of amblyopia (best-corrected visual acuity of 20/40 or worse at presentation) as a result of treatment also showed significant treatment effects: Most astigmatic children who met the criterion for amblyopia at baseline no longer met the criterion after 1 year of treatment. However, based on the percentage of nonastigmats who met the criterion for amblyopia at baseline (12.0% in the YC and 4.3% in the OC), approximately 12% of younger astigmats and 4% of older astigmats may have been overdiagnosed, because children in the nonastigmatic group did not have any ocular abnormalities that might have led to amblyopia. Most likely, these children performed poorly on the visual acuity test for reasons unrelated to visual performance (e.g., some children, particularly younger children, may have had difficulty with the task or were distracted during the task). This is supported by the fact that only 1 of the nonastigmatic children who met the criterion for amblyopia at baseline still met the criterion at 1 year, despite the fact that no treatment was administered. These results raise the possibility that the visual acuity criterion for amblyopia diagnosis (20/40 or worse) may have been too rigorous, particularly for the YC, and that some of the astigmatic children might have been overdiagnosed as amblyopes at baseline, leading to an overestimation of the percentage of children successfully treated (i.e., percentage who no longer met the criterion for amblyopia). However, if we look more conservatively at our treatment effects by reducing the percentage of children in the YC who were treated successfully by 12% and the percentage in the OC by 4%, the percentage of children who were treated successfully remains substantial (66% and 48%, respectively, for previously untreated and previously treated in the YC; 96% and 64%, respectively, for previously untreated and previously treated in the OC). The results of the present study suggest that the introduction of clear visual input (through eyeglass correction of astigmatic refractive error) is sufficient to result in an improvement in best-corrected visual acuity in school-age children with astigmatism-related amblyopia and that this treatment effect occurs over a period of approximately 6 weeks. The results also indicate that the sensitive period for treatment of astigmatism-related amblyopia with eyeglass correction alone extends beyond age 7 years, contrary to previous findings in the literature on astigmatism-related amblyopia. 6,12,14 The results of the present study are not consistent with the results of our previous study, in which we did not find significant treatment effects as a result of eyeglass correction of astigmatism in 3- to 5-year-old children. 14 A factor that may have contributed to the difference in effects between our earlier study of preschool children and the present study is the difference in procedures used in the 2 studies. In the preschool study, children were tested, both at baseline and at follow-up, after they had undergone cycloplegic refraction and while they were wearing trial frames. In contrast, children in the present study were tested while wearing eyeglasses and were not cyclopleged during vision testing. Thus, variability produced by testing under cycloplegia and with trial frames may have masked any improvement in best-corrected visual acuity that occurred between baseline and follow-up testing in the preschool children. The results of the present study are consistent with those of previous studies that have demonstrated a significant response to eyeglass treatment alone in children with anisometropic or strabismic amblyopia, or both For example, Moseley et al 24 reported improvement in all 13 ambly- 2299

8 Ophthalmology Volume 114, Number 12, December 2007 opic subjects (age range, 3 6 years) treated with eyeglass correction alone, 5 of whom had significant astigmatism ( 1.00 D). Stewart et al 25 reported a treatment response to eyeglass correction alone in anisometropic and strabismic amblyopes, with an average of 14 weeks to best bestcorrected visual acuity and equal gains at all ages between age 3 and 8 years. In a randomized trial of eyeglass correction only versus eyeglass correction plus patching (atropine also prescribed for younger [7- to 12-year-old] children) in 7- to 17-year-old children with strabismic or anisometropic amblyopia, or both, the Pediatric Eye Disease Investigator Group reported significant improvement with eyeglass correction alone. 26 Furthermore, response to eyeglass treatment did not seem to be dependent on age. That is, in 7- to 12-year-old children, 25% of the eyeglass correction group were considered treatment responders (10 or more letters improvement in visual acuity, relative to baseline acuity), and in 13- to 17-year-old children, 23% of the eyeglass correction group were considered treatment responders. Treatment Effects Occur Early in Treatment Results of the present study suggest that treatment effects occur primarily early in treatment (within the first few weeks; average, 6 weeks), because there were no significant treatment effects observed from the 6-week to the 1-year follow-up. Stewart et al 25 reported an average of 1 line of improvement in acuity in the amblyopic eye in anisometropic and strabismic amblyopes after 6 weeks of eyeglass correction alone and 2 lines improvement after 12 weeks, with maximum improvement shown after approximately 15 weeks. In a recent study of previously untreated 3- to 6-year-old children with anisometropic amblyopia, the Pediatric Eye Disease Investigator Group reported an average acuity improvement of 2.9 lines with eyeglass correction alone, with most improvement occurring within the first 5 weeks (mean improvement, 1.8 lines). 27 In most subjects, improvement occurred within the first few months: 83% stopped improving within 15 weeks of initiating eyeglass wear. However, the study also provided evidence to suggest that further improvement would have occurred in some subjects if they had been followed up for a longer period. If no improvement occurred within a 5-week follow-up interval, subjects were classified as experiencing no further improvement and were randomized to other treatments. Results from children who then were randomized into the eyeglass control group indicated that there was an average of 1.2 lines of further improvement. It is possible that improvement occurs more slowly after the initial treatment period and that 5 weeks is not a sufficient duration to detect smaller changes. However, the insufficient follow-up interval explanation for null treatment effects beyond the first 6 weeks is less likely in the present study because subjects were followed up for an average of 1 year. Residual Deficits: Poor Compliance, Reduced Plasticity, or Uncorrected Refractive Errors? Although we did find significant improvement with treatment, the results also indicated that, after 1 year of eyeglass treatment, the astigmatic groups did not reach the acuity level of the nonastigmatic control group (Fig 2). This finding can be interpreted in 3 ways: (1) plasticity was limited in this age range and subjects did not have the capacity to achieve normal levels of acuity; (2) although there were treatment effects in compliant subjects, there were null or reduced treatment effects in poorly compliant subjects, resulting in a reduction in the overall (mean) visual performance in astigmatic children compared with nonastigmatic children; (3) there were small residual refractive errors, for example, spherical aberrations or coma, that were present but not corrected with eyeglasses in the astigmatic groups, and this failure to provide full correction limited the improvement seen in the astigmatic groups over time; or a combination thereof. Thus, the analyses of treatment effects reflect effectiveness (results of all treated, regardless of compliance), rather than efficacy (results of treatment in ideal (perfect compliance) circumstances). Secondary analyses of best-corrected visual acuity data for compliant astigmatic subjects did not support the hypothesis that poor treatment compliance contributed to the failure of the astigmatic group to reach normal (nonastigmatic group) levels of best-corrected visual acuity: there was no evidence of greater treatment effects in compliant than in noncompliant subjects, nor did compliant subjects reach, on average, the best-corrected visual acuity level of the nonastigmatic group after 1 year of treatment. However, we can not rule out the possibility that with perfect compliance and sufficient treatment duration, astigmatic subjects might have reached normal best-corrected visual acuity levels. At this point, we can not rule out the hypothesis that there were residual (uncorrected) refractive errors in astigmats (e.g., spherical aberrations or coma) that hindered the improvement resulting from eyeglass wear. An ongoing study of refractive error development and visual development in Tohono O odham infants and young children in which we are longitudinally measuring spherical aberrations and coma will provide more definitive answers as to whether there is an association between astigmatism and higher-order aberrations. Strengths and Limitations An important strength of the study design is the inclusion of a nonastigmatic control group as a normal reference to allow us to determine if best-corrected visual performance was significantly below normal in the treatment groups at baseline and after treatment. In addition, comparison with a normal control group allows us to rule out the possibility that improvements over time in the astigmatic groups were the result of practice effects, cognitive changes in test performance ability, or normal improvement in visual function resulting from development. A weakness of the present study is the absence of an untreated astigmatic control group that was followed up prospectively. We chose to include a nonastigmatic control group, rather than an untreated astigmatic control group, to avoid withholding eyeglass correction to the astigmatic children enrolled in the study. We believed that this was 2300

9 Harvey et al Treatment of Astigmatism-Related Amblyopia ethically appropriate for 2 reasons: (1) because eyeglass correction of astigmatism provides a significant improvement in visual acuity to astigmatic children, even in the presence of amblyopia, 14 and (2) because we did not want to withhold treatment that was likely to become less effective if withheld, because the available research indicated that astigmatism-related amblyopia becomes less responsive to treatment as children grow older. 6,12,14 Therefore, the possibility that the astigmatism groups would have improved more than the nonastigmatic group over time with no treatment at all can not be ruled out. However, this is unlikely for 2 reasons: (1) the observed treatment effect occurred within just a few weeks of treatment initiation, that is, during a time frame in which natural maturation is not likely to have had a significant effect; and (2) if astigmatic children improved over time regardless of eyeglass treatment, the previously untreated OC should have demonstrated lesser deficits at baseline than the previously untreated YC, and this was not the case (the interaction between group and cohort was not significant at baseline). Conclusions The results of the present study provide strong evidence that children older than age 7 years do respond to eyeglass treatment of astigmatism-related amblyopia and suggest that eyeglass treatment alone may be associated with significant reduction in amblyopia in children with bilateral ametropic amblyopia. These findings may provide incentive for practitioners to stress more strongly to their school-age patients the importance of eyeglass wear for the treatment of astigmatismrelated amblyopia. The present study and other recent studies showing responsiveness to amblyopia treatment in older children also should provide incentive for further investigation into methods of amblyopia treatment in children beyond the preschool and early elementary school years. References 1. Atkinson J, Braddick O, Bobier B, et al. Two infant vision screening programmes: prediction and prevention of strabismus and amblyopia from photo- and videorefractive screening. Eye 1996;10: Dobson V, Tyszko RM, Miller JM, Harvey EM. Astigmatism, amblyopia, and visual disability among a Native American population. In: Vision Science and Its Applications: Summaries of the Papers Presented at the Topical Meeting, February 1 5, 1996, Santa Fe, New Mexico. Washington: Optical Society of America; 1996: OSA Technical Digest Series. Vol Dobson V, Miller JM, Harvey EM, Mohan KM. Amblyopia in astigmatic preschool children. Vision Res 2003;43: Harvey EM, Dobson V, Miller JM, Clifford-Donaldson CE. Amblyopia in astigmatic children: patterns of deficits. Vision Res 2007;47: Freeman RD, Mitchell DE, Millodot M. A neural effect of partial visual deprivation in humans. Science 1972;175: Mitchell DE, Freeman RD, Millodot M, Haegerstrom G. Meridional amblyopia: evidence for modification of the human visual system by early visual experience. Vision Res 1973;13: Mohindra I, Jacobson SG, Held R. Binocular visual form deprivation in human infants. Doc Ophthalmol 1983;55: Mitchell DE, Wilkinson F. The effect of early astigmatism on the visual resolution of gratings. J Physiol 1974;243: Freeman RD. Asymmetries in human accommodation and visual experience. Vision Res 1975;15: Freeman RD. Contrast sensitivity in meridional amblyopia. Invest Ophthalmol 1975;14: Freeman RD, Thibos LN. Contrast sensitivity in humans with abnormal visual experience. J Physiol 1975;247: Cobb SR, MacDonald CF. Resolution acuity in astigmats: evidence for a critical period in the human visual system. Br J Physiol Opt 1978;32: Gwiazda J, Scheiman M, Held R. Anisotropic resolution in children s vision. Vision Res 1984;24: Harvey EM, Dobson V, Miller JM, Sherrill DL. Treatment of astigmatism-related amblyopia in 3- to 5-year-old children. Vision Res 2004;44: Dobson V, Miller JM, Harvey EM. Corneal and refractive astigmatism in a sample of 3- to 5-year-old children with a high prevalence of astigmatism. Optom Vis Sci 1999;76: Dobson V, Miller JM, Harvey EM, Sherrill DL. Prevalence of astigmatism, astigmatic anisometropia, and glasses wearing among preschool- and school-age Native American children. In: Vision Science and Its Applications: February 19 22, 1999, Santa Fe, New Mexico. Washington: Optical Society of America; 1999: OSA Technical Digest Series. 17. Harvey EM, Dobson V, Miller JM. Prevalence of high astigmatism, eyeglass wear, and poor visual acuity among Native American grade school children. Optom Vis Sci 2006;83: Miller JM, Dobson V, Harvey EM, Sherrill DL. Astigmatism and Amblyopia Among Native American Children (AANAC): design and methods. Ophthalmic Epidemiol 2000;7: Miller JM, Dobson V, Harvey EM, Sherrill DL. Comparison of preschool vision screening methods in a population with a high prevalence of astigmatism. Invest Ophthalmol Vis Sci 2001;42: Long WF. A matrix formalism for decentration problems. Am J Optom Physiol Opt 1976;53: Harris WF. The mean and variance of samples of dioptric powers: the basic calculations. Clin Exp Optom 1990;73: Guyton DL, Miller JM, West CE. Optical pearls and pitfalls. In: Wright KW, Spiegel PH, eds. Pediatric Ophthalmology and Strabismus. 2nd ed. New York: Springer; 2003: Ferris FL III, Kassoff A, Bresnick GH, Bailey I. New visual acuity charts for clinical research. Am J Ophthalmol 1982;94: Moseley MJ, Neufeld M, McCarry B, et al. Remediation of refractive amblyopia by optical correction alone. Ophthalmic Physiol Opt 2002;22: Stewart CE, Moseley MJ, Fielder AR, et al. Refractive adaptation in amblyopia: quantification of effect and implications for practice. Br J Ophthalmol 2004;88: Pediatric Eye Disease Investigator Group. Randomized trial of treatment of amblyopia in children aged 7 to 17 years. Arch Ophthalmol 2005;123: Pediatric Eye Disease Investigator Group. Treatment of anisometropic amblyopia in children with refractive correction. Ophthalmology 2006;113: