Although there is agreement that amblyopia can be treated successfully in most children, there is uncertainty regarding the long-term outcome in visual acuity after cessation of treatment. 1–7 Also, the natural evolution of refractive error for these patients is not well documented. Parents frequently ask whether the improvement in visual acuity gained with treatment will last and whether their children will always need an optical correction.
Knowledge about the development of visual acuity and refractive error in amblyopes several years after cessation of treatment is limited. A small number of studies exist on the durability of amblyopia treatment. The consensus is that many children treated for either anisometropic or strabismic amblyopia regress in visual acuity during their adolescent and teenage years. Rutstein and Fuhr 2 reported that approximately 70% lose, on average, 43% of the visual gain in their amblyopic eye within 1 year after therapy cessation. Others found that the visual acuity gains endured in 5% to 78% of their patients. 3–7 Recently, Ohlsson and associates 8 reported that visual acuity deteriorated in 17% of their patients, remained stable in 50%, and improved in 33% after a mean follow-up of 10.4 years. No patient lost more than 2 lines of visual acuity.
The type of amblyopia may determine its course. Strabismic amblyopes may regress in visual acuity more than anisometropic amblyopes because of the former's lack of achieving normal binocular vision and stereopsis after treatment. 9 The type of treatment may also determine the durability of the visual acuity gains, at least for anisometropic amblyopia. One study found that visual acuity in the amblyopic eye 1 to 2 years after cessation of therapy endured in 50% of anisometropic amblyopes treated with optical correction only; in 60% treated with optical correction and occlusion; and in 100% treated with optical correction, occlusion, and vision therapy. 10
Changes in the refractive error for amblyopic and nonamblyopic eyes over the years have been reported as not being symmetric. Lepard 11 reported that the nonamblyopic, or fixating, eye in strabismics became much less hyperopic or much more myopic, whereas no appreciable changes or slightly increased hyperopia occurred in the amblyopic eye. The average interval between the first and the last refraction was 14 years. Other investigators 12 reported diminishing hyperopia and/or increasing myopia in both the nonamblyopic and amblyopic eyes, with the nonamblyopic eye having much greater changes. In addition, the amount of anisometropia in nonstrabismic amblyopes has been shown to decrease over time. 8 Changes in refractive error in the amblyopic and the nonamblyopic eye may be related to whether either hyperopia or myopia is present initially. Greater changes have been reported to occur in the nonamblyopic eye versus the amblyopic eye for strabismic patients who are hyperopic, whereas greater changes have been reported to occur in the amblyopic eye for strabismic patients who are myopic. 13
The purpose of this study was to evaluate the long-term outcome in visual acuity and refractive error in both amblyopic and nonamblyopic eyes in patients treated in the clinic.
After approval by the University of Alabama at Birmingham Institutional Review Board, the records of all children treated for strabismic, anisometropic, and isoametropic amblyopia in the school's clinic from 1983 to 1993 were reviewed. Treatment, in addition to spectacles, included mostly adhesive occlusion of the nonamblyopic eye in strabismic amblyopia and anisometropic amblyopia. The amount of occlusion varied from all waking hours to 3 to 4 hours per day. One patient was treated with crossed polarizing lenses, 14 and another was treated with optical penalization. 15 For isoametropic amblyopia, treatment consisted of spectacle correction. Amblyopia was defined as an interocular difference in initial best-correctable visual acuity of 1 line or more and the best-correctable visual acuity of the amblyopic eye of 20/30 or poorer (strabismic and anisometropic amblyopia) or best-correctable visual acuity of 20/30 or poorer in each eye for isoametropic amblyopia. In addition, an amblyopiagenic factor (strabismus, an-isometropia, or high refractive error) had to be present. A total of 320 records was identified.
Excluded were patients who had fewer than 4 years of follow-up after cessation of amblyopia treatment and patients with deprivation amblyopia, neurological impairment, systemic disease, developmental delay, ocular disease, or disease of the visual pathways. Also excluded were patients who at any clinic visit were incapable of responding verbally to visual acuity testing, as well as patients who did not comply with amblyopia treatment (by their parents or their own report).
Based on these criteria, records of 61 patients having strabismic amblyopia, anisometropic amblyopia, or isoametropic amblyopia were included in the study. Anisometropia was defined as a difference of ≥1.50 D interocularly in either the spherical or cylindrical component of the refraction. Patients with anisometropic amblyopia were nonstrabismic at both distance and near and fixated centrally. Patients with strabismus in combination with anisometropia were classified as having strabismic amblyopia. Isoametropia was defined as a bilateral refractive error of ≥3 D with <1.50 D interocular difference in any meridian. Patients with isoametropic amblyopia were nonstrabismic.
The following information was obtained from each record: (1) age of patient when amblyopia treatment was initiated; (2) patient gender; (3) length of follow-up after cessation of treatment; (4) type of amblyopia; (5) ocular alignment status; (6) initial best-correctable visual acuity in amblyopic and nonamblyopic eyes (pretreatment); (7) best-correctable visual acuity in amblyopic and nonamblyopic eyes at the conclusion of amblyopia treatment for strabismic amblyopia and anisometropic amblyopia (posttreatment); (8) best-correctable visual acuity in amblyopic and nonamblyopic eyes at the final clinic visit (follow-up); (9) the refractive error in each eye at both pretreatment and follow-up visits; and (10) stereoacuity at the pretreatment visit.
During the treatment period, visual acuity was assessed for each patient with tests appropriate for the patient's age. In all cases, the method of measurement was consistent during the treatment period. At the follow-up visit, Snellen chart (line acuity) was the most common method of visual acuity measurement. For analysis, all visual acuity values were recalculated to logMAR units (the logarithm of the Snellen reciprocal). Stereoacuity was measured at near with either the Titmus or Randot stereo tests. Ocular alignment status was evaluated at distance and near with the unilateral cover test. Only records wherein the same method of refraction, cycloplegic or noncycloplegic, was performed at the pretreatment visit and follow-up visit were included. All refractive errors were converted to spherical equivalents.
The data in Table 1 list the number of patients having each type of amblyopia and the minimum, mean, and maximum ages at the pretreatment visit. Patients with strabismic amblyopia tended to be treated at a significantly younger age than those with isoametropic amblyopia but not those with anisometropic amblyopia (F2,58 = 5.53, p = 0.0063). Table 1 also shows the time in years between visits for each type of amblyopia. Patients with strabismic amblyopia also tended to have longer follow-up, but this was not statistically different from those with anisometropic or isoametropic amblyopia.
Table 2 lists the mean changes in logMAR visual acuity for strabismic amblyopia and anisometropic amblyopia. The mean Snellen visual acuity in the amblyopic eye for strabismic amblyopia before treatment was approximately 20/80, whereas the mean Snellen visual acuity for anisometropic amblyopia was approximately 20/63. Treatment improved visual acuity in the amblyopic eye for strabismic amblyopia 0.36 logMAR unit (approximately 3.5 Snellen lines) and for anisometropic amblyopia 0.30 logMAR unit (approximately 3 Snellen lines) after an average treatment period of 1 and 1.1 years, respectively. At the final visit, 9.3 years after cessation of treatment for strabismic amblyopia and 7.1 years after cessation of treatment for anisometropic amblyopia, the mean visual acuity regressed 0.09 logMAR unit (approximately 1 Snellen line) for both strabismic amblyopia and anisometropic amblyopia. Repeated-measures analysis of variance (ANOVA) showed a significant effect of visit (F2,92 = 36.8, p < 0.0001) on visual acuity in the amblyopic eye, independent of the type of amblyopia. A post hoc comparison of the means (unequal n, honestly significant difference test) 16 for the three visits showed that the effect was due to the changes between the pretreatment and posttreatment visits and that the changes between the posttreatment and follow-up visits were not significant (p = 0.07).
For strabismic amblyopia, 12 patients (55%) maintained visual acuity in the amblyopic eye equal to or within 0.1 logMAR unit, or 1 Snellen line, of that gained by treatment; eight patients (36%) had their visual acuity regress by >0.1 logMAR unit, and two patients (9%) had their visual acuity improve by >0.1 logMAR unit at the follow-up visit. Five of 19 patients (26%) tested had measurable stereopsis at the pretreatment visit (mean = 210 s; range, 50 to 400 s). Four of these patients (80%) either maintained or gained visual acuity.
For anisometropic amblyopia, 12 patients (46%) maintained visual acuity in the amblyopic eye within 0.1 logMAR unit, or 1 Snellen line, of their best visual acuity after treatment; nine patients (35%) had their visual acuity regress by >0.1 logMAR unit, and five patients (19%) had their visual acuity improve by >0.1 logMAR unit. Twenty of 26 patients (77%) tested had measurable stereopsis at the pretreatment visit (mean = 109 s; range, 20 to 80 s). Fifteen (75%) either maintained or gained visual acuity, and 5 patients (25%) lost visual acuity.
Table 2 indicates that the mean visual acuity in the nonamblyopic eye for both strabismic amblyopia and anisometropic amblyopia also improved from the pretreatment to the follow-up visit. The mean visual acuity in the nonamblyopic eye for strabismic amblyopia was approximately 20/30 before treatment, whereas the mean visual acuity in the nonamblyopic eye for anisometropic amblyopia was approximately 20/25. Repeated-measures ANOVA show that the changes in visual acuity in the nonamblyopic eye with visit were significant (F2,92 = 28.3, p = 0.0001) and interacted with type (F2,92 = 6.7, p = 0.0019), the changes being larger in strabismic amblyopia.
Figure 1 illustrates the magnitude of the combined changes in logMAR visual acuity for amblyopic and nonamblyopic eyes between visits for strabismic amblyopia and anisometropic amblyopia combined. Although the post hoc comparison of the absolute values of the mean logMAR acuities showed that the mean absolute value of visual acuity at the follow-up visit was not significantly different from the mean visual acuity at the posttreatment visit, the magnitude of the changes in visual acuity was different in a number of ways. In the amblyopic eye, the magnitude of the changes in visual acuity between the pretreatment and posttreatment visits was significantly different from those that occurred between the posttreatment and follow-up visits (t = 6.29, p < 0.0001). The positive change in the latter is indicative of regression of visual acuity. However, the magnitude of change in visual acuity from the pretreatment to the posttreatment visit was not significantly different from that between the pretreatment to the follow-up visit, indicating that the regression did not significantly reduce the gains made by treatment.
In the nonamblyopic eye, there was no significant difference in the magnitude of the changes that occurred between pretreatment and posttreatment visits and between posttreatment and follow-up visits. When the magnitude of these changes was compared with that for the amblyopic eye, it is clear, however, that the regression in visual acuity in the amblyopic eye represents a significant effect when compared with the changes in the normal or nonamblyopic eye (t = 3.81, p = 0.0004). The latter continued to improve while the former actually became worse.
Table 3 lists the spherical equivalent refractive errors for strabismic amblyopia and anisometropic amblyopia. In strabismic amblyopia, the mean refractive error in amblyopic and nonamblyopic eyes changed from +2.15 D and +1.85, respectively, to +0.45 D and +0.58 D, respectively, between pretreatment and follow-up visits. In anisometropic amblyopia, the mean refractive error in amblyopic and nonamblyopic eyes changed from +1.04 D and +0.12, respectively, to +0.23 D and −0.94 D, respectively, between pretreatment and follow-up visits. The effect of visit on both the amblyopic and the nonamblyopic refractive error was significant (F1,46 = 36.7, p < 0.0001 and F1,46 = 29.8, p < 0.0001, respectively).
Significant improvement in visual acuity for both eyes in patients corrected for isoametropic amblyopia occurred after a mean follow-up of 8.9 years (Table 4). The mean visual acuity in the right and the left eye at the pretreatment visit was approximately 20/50. The average change for both eyes was 0.25 logMAR unit (approximately 2.5 Snellen lines). Only 1 patient did not improve. This patient had minimal amblyopia in each eye (20/30) at the pretreatment visit. Three patients (25%) achieved a visual acuity of 0.0 logMAR unit (20/20). Eleven patients (92%) tested had measurable stereopsis at the pretreatment visit (mean = 81.8 s; range, 20 to 200 s). Nine patients (75%) had a myopic refractive error at the pretreatment visit. There was a significant myopic shift in each eye during follow-up (Table 5).
The ultimate goal of amblyopia treatment is to improve visual acuity and prevent its recurrence. 17 For the patients in the present study, the mean gain in visual acuity after treatment was similar for strabismic amblyopia and anisometropic amblyopia. The mean regression after 9.3 years after cessation of treatment for strabismic amblyopia and after 7.1 years after cessation of treatment for anisometropic amblyopia was also similar and statistically insignificant. As many as 65% either maintained or gained visual acuity in the amblyopic eye over the years. This retention rate is somewhat better than the 53% reported by Flynn and associates 6 after a mean follow-up of 4.6 years, similar to the 67% reported by Leiba and associates 7 after a mean follow-up of 21.5 years, and less than the 83% reported by Ohlsson and associates 8 after a mean follow-up of 10.4 years.
In a prior study, nearly 70% of amblyopes lost, on average, 43% of their visual gain within 1 year after treatment completion. 2 Unlike the present patients who had a mean age at follow-up of approximately 16 years, most of the patients from the earlier study were 8 years or younger at follow-up. With longer follow-up and when the patient is visually mature, visual acuity gained by treatment appears to be maintained more readily, and the risk for significant regression becomes less likely. 7, 8 Levartovsky et al. 9 found that amblyopic children with occlusion treatment and follow-up monitoring up to 9 years of age had less deterioration of visual acuity in the amblyopic eye at long-term follow-up than did children whose monitoring ended before age 9.
The mean visual acuity in the nonamblyopic eye also improved, more so for strabismic amblyopia than for anisometropic amblyopia. These changes were statistically significant from the pretreatment to the follow-up visit. Olhsson and associates 8 also reported statistically significant visual acuity improvement in the nonamblyopic eye for their patients. The mean improvement that they reported, 0.06 logMAR unit (approximately half a line), is similar to the improvement for our patients with anisometropic amblyopia.
Some investigators have reported that the nonamblyopic eye in untreated adult amblyopes has subnormal visual performance and does not achieve a normal age-corresponding physiological status and normal visual acuity. 18–20 This suggests that the improvement in visual acuity in the nonamblyopic eye in the treated patients in the present study is real. Another possibility is that the improvement in the nonamblyopic eye is due partly to a learning or an age effect. The average magnitude of any potential learning and age effect on visual acuity in a recent study comparing minimal versus part-time occlusion in amblyopic children after 4 months of treatment was 0.14 line. 21 For the patients in the present study, the combined average improvement in visual acuity from the pretreatment to the follow-up visit in the nonamblyopic eye for strabismic amblyopia and anisometropic amblyopia was greater, 0.1 logMAR unit, or approximately 1 Snellen line (Fig. 1).
The mean refractive error became significantly less hyperopic and/or significantly more myopic in both nonamblyopic and amblyopic eyes in strabismic amblyopia and anisometropic amblyopia. This finding contradicts earlier studies that showed that the eye with normal visual acuity in strabismic amblyopia either emmetropizes or becomes significantly myopic with age, whereas the amblyopic eye fails to emmetropize and remains the same, becomes only slightly less hyperopic, or becomes slightly more hyperopic. 11, 12 Unlike those in previous studies, the present patients had a shorter follow-up period between the first and last refraction and included both anisometropic amblyopia and strabismic amblyopia. The present patients also had, on average, smaller amounts of baseline hyperopia at the pretreatment visit and slightly better visual acuity in the amblyopic eye at the follow-up visit. The lower hyperopia is likely due to the fact that some patients received noncycloplegic refractions at the pretreatment visit.
The type of refractive error at the pretreatment visit may determine whether the amblyopic or the nonamblyopic eye manifests the greater refractive change over time. Burtolo and associates 13 reported the changes in refractive error in hyperopic and myopic children (mean age, 5.1 years) with strabismic amblyopia after a follow-up of 3 years. For the hyperopes, the refractive error in the nonamblyopic eye became less hyperopic by an average of 1.30 D, whereas the amblyopic eye became less hyperopic by an average of 0.40 D. For the myopes, the nonamblyopic eye became more myopic by 0.50 D, and the amblyopic eye became more myopic by 2.80 D. Burtolo and associates 13 confirmed these changes by echo-biometry, which showed corresponding axial-length changes: less elongation in the amblyopic eyes of hyperopes and more elongation in the amblyopic eyes of myopes. For the patients in the present study, the mean refractive error at the pretreatment visit in amblyopic and nonamblyopic eyes for strabismic amblyopia was +2.15 D and +1.85 D, respectively, whereas the mean refractive error for anisometropic amblyopia was +1.04 D and +0.12 D, respectively. A slightly larger myopic shift occurred in the amblyopic eye (1.70 vs. 1.27 D) in strabismic amblyopia, whereas in anisometropic amblyopia, a slightly larger myopic shift occurred in the nonamblyopic eye (0.81 vs. 1.06 D). The average amount of anisometropia decreased in strabismic amblyopia (0.30 to 0.13 D) and increased in anisometropic amblyopia (0.92 to 1.17 D).
For patients with isoametropic amblyopia, visual acuity improved symmetrically in both eyes with the appropriate spectacle correction. Unlike that in prior studies, 22, 23 the mean refractive error for patients with isoametropic amblyopia at the pretreatment visit was myopia and not hyperopia. Patients with isoametropic amblyopia also tended to manifest a myopic shift bilaterally over the years.
There are limitations to the present study. The inclusion criteria specified only using records of patients who had been fully compliant during the treatment period and who had a follow-up ≥4 years after cessation of treatment. Being retrospective, the measurement of visual acuity was not identical for every patient. Different examiners and different acuity charts were utilized. However, all visual acuities were measured with the same method during the treatment period. At the follow-up visit, most visual acuities were measured with the Snellen chart. Also, the method of refraction was not similar for the patients, some having cycloplegic and some having noncycloplegic refractions. However, only records of patients for whom the same type of refraction was performed at the pretreatment and follow-up visits were included.
In summary, our findings suggest that after treatment and with long-term follow-up, visual acuity regresses on average but not significantly for most patients with both strabismic amblyopia and anisometropic amblyopia. Visual acuity also seems to improve slightly in the nonamblyopic eye, more so for strabismic amblyopia. This may be due in part to an age and a learning effect. Visual acuity also improves over time in isoametropic amblyopia. The refractive error of both eyes tends to show a myopic shift, regardless of the type of amblyopia.
Presented at the annual meeting of the Association for Research and Vision in Ophthalmology on May 8, 2003, in Ft. Lauderdale, Florida.
1. Birnbaum MH, Koslowe K, Sanet R. Success in amblyopia therapy as a function of age: a literature survey. Am J Optom Physiol Opt 1977;54:269–75.
2. Rutstein RP, Fuhr PS. Efficacy and stability of amblyopia therapy. Optom Vis Sci 1992;69:747–54.
3. Leydhecker W, Ricklefs G, Ruhling R. Late results of amblyopia treatment. Klin Monatsbl Augenheilkd 1967;151:373–6.
4. Ching FC, Parks MM, Friendly DS. Practical management of amblyopia. J Pediatr Ophthalmol Strabismus 1986;23:12–6.
5. Scott WE, Dickey CF. Stability of visual acuity in amblyopic patients after visual maturity. Graefes Arch Clin Exp Ophthalmol 1988;226:154–7.
6. Flynn JT, Schiffman J, Feuer W, Corona A. The therapy of amblyopia: an analysis of the results of amblyopia therapy utilizing the pooled data of published studies. Trans Am Ophthalmol Soc 1998;96:431–50.
7. Leiba H, Shimshoni M, Oliver M, Gottesman N, Levartovsky S. Long-term follow-up of occlusion therapy in amblyopia. Ophthalmology 2001;108:1552–5.
8. Ohlsson J, Baumann M, Sjostrand J, Abrahamsson M. Long term visual outcome in amblyopia treatment. Br J Ophthalmol 2002;86:1148–51.
9. Levartovsky S, Oliver M, Gottesman N, Shimshoni M. Factors affecting long term results of successfully treated amblyopia: initial visual acuity and type of amblyopia. Br J Ophthalmol 1995;79:225–8.
10. FitzGerald DE, Krumholtz I. Maintenance of improvement gains in refractive amblyopia: a comparison of treatment modalities. Optometry 2002;73:153–9.
11. Lepard CW. Comparative changes in the error of refraction between fixing and amblyopic eyes during growth and development. Am J Ophthalmol 1975;80:485–90.
12. Nastri G, Perugini GC, Savastano S, Polzella A, Sbordone G. The evolution of refraction in the fixing and the amblyopic eye. Doc Ophthalmol 1984;56:265–74.
13. Burtolo C, Ciurlo C, Polizzi A, Lantier PB, Calabria G. Echobiometric study of ocular growth in patients with amblyopia. J Pediatr Ophthalmol Strabismus 2002;39:209–14.
14. Wesson MD. Use of light intensity reduction for amblyopia therapy. Am J Optom Physiol Opt 1983;60:112–7.
15. Repka MX, Ray JM. The efficacy of optical and pharmacological penalization. Ophthalmology 1993;100:769–74.
16. Winer BJ, Brown DR, Michels KM. Statistical Principals in Experimental Design, 3rd ed. New York: McGraw-Hill, 1991.
17. Oster JG, Simon JW, Jenkins P. When is it safe to stop patching? Br J Ophthalmol 1990;74:709–11.
18. Hasse W, Wenzel F. The natural course of untreated functional amblyopia: does it progress between childhood and adulthood? Binoc Vis Strabismus Q 1997;12:17–24.
19. Simons K, Preslan M. Natural history of amblyopia untreated owing to lack of compliance. Br J Ophthalmol 1999;83:582–7.
20. Leguire LE, Rogers GL, Bremer DL. Amblyopia: the normal eye is not normal. J Pediatr Ophthalmol Strabismus 1990;27:32–8.
21. Repka MX, Beck RW, Holmes JM, Birch EE, Chandler DL, Cotter SA, Hertle RW, Kraker RT, Moke PS, Quinn GE, Scheiman MM. A randomized trial of patching regimens for treatment of moderate amblyopia in children. Arch Ophthalmol 2003;121:603–11.
22. Fern KD. Visual acuity outcome in isometropic hyperopia. Optom Vis Sci 1989;66:649–58.
© 2004 American Academy of Optometry
23. Werner DB, Scott WE. Amblyopia case reports: bilateral hypermetropic ametropic amblyopia. J Pediatr Ophthalmol Strabismus 1985;22:203–5.