Occlusion or iatrogenic amblyopia is characterized by a decrease in visual acuity that can develop during amblyopia treatment in the occluded or penalized eye. The formerly amblyopic eye becomes the dominant eye. Occlusion amblyopia is often associated with full-time patching.1–5 Although occlusion amblyopia is well documented, few large studies have specifically investigated this condition. The purpose of this article is to determine the incidence, patient characteristics, and long-term visual acuity outcomes associated with occlusion amblyopia following full-time patching therapy.
Patients and Methods
A retrospective chart review was performed after approval by the Institutional Review Board for patients with the diagnosis of amblyopia seen at the University of Iowa Hospitals and Clinics Pediatric Ophthalmology Clinic between 1970 and 2000. Only patients who were prescribed full-time patching for amblyopia were included. Patients were excluded because they were not treated with full-time occlusion, had no patching, were older than 10 years when diagnosed as having amblyopia, were lost to follow-up, had unavailable records, or had amblyopia from causes other than strabismus, anisometropia, or both. Patients with structural ocular abnormalities were also excluded.
Full-time patching was defined as 24 hours per day of patching or patching of the fixing eye during all waking hours.6 All patients were observed until the completion of full-time patching. The endpoint of full-time patching was either equal vision between the two eyes or until no improvement occurred after three age-related episodes of full-time patching with good compliance.7 An age-related episode was defined as 1 week of patching per year of age up to a maximum of 4 weeks. For example, an age-related episode for a 2-year-old child would be 2 weeks.
Strabismus was measured with single and alternate prism cover testing at 33 cm and 6 m while wearing proper spectacle correction on an accommodative target. Those classified as having amblyopia due to strabismus had measurements of 8 or more prism diopters of horizontal deviation or 5 or more prism diopters of vertical deviation or a history of strabismus surgery. Those classified as having amblyopia due to anisometropia had a refractive difference of 1 diopter or more in any meridian. Combination amblyopia was defined as any patient who met the criteria for both strabismic and anisometropic amblyopia.
Occlusion amblyopia was defined as a decrease in optotype visual acuity of the originally fixing eye to two or more lines worse than the visual acuity of the originally amblyopic eye at a single follow-up visit during the course of full-time occlusion treatment. Occlusion amblyopia was not defined as a decrease from an original baseline visual acuity of the fixing eye because the baseline visual acuity of children was often noted to improve in both eyes during amblyopia treatment. Using this definition could result in the underreporting of occlusion amblyopia. A switch in interocular acuity, with the previously amblyopic eye having better vision relative to the originally fixing eye, was believed to be a more accurate indicator of occlusion amblyopia. In patients too young to perform optotype visual acuity testing, occlusion amblyopia was defined as a change in fixation in the occluded eye from maintained to not maintained by fixation preference testing. Occlusion amblyopia was treated with a cessation of all patching, reverse occlusion, or alternate patching.
The ages at onset of full-time patching and occlusion amblyopia, the method of visual acuity testing used, and the number of age-related episodes of full-time patching that occurred prior to the onset of occlusion amblyopia were documented. The best-corrected visual acuity in both eyes at initial examination and final examination was recorded and the final visual acuity difference in lines between the two eyes at least 6 months after stopping all patching therapy was determined after converting optotype visual acuities to their logarithm of the minimum angle of resolution equivalent.
The frequency count and percentage (and 95% confidence limits) with occlusion amblyopia was calculated for each full-time occlusion episode. The generalized linear model analysis was used to fit a multi-factor logistic regression model to identify factors associated with occlusion amblyopia, which included type of amblyopia, age at first episode of full-time occlusion, method of testing, and number of full-time occlusion episodes. Because a patient is at risk for occlusion amblyopia at each full-time occlusion episode, to be able to test for the effect of full-time occlusion episode, each full-time occlusion episode for a patient was considered as a binary event with the outcome of either yes or no for occlusion amblyopia. For example, patients who had two full-time occlusion episodes and developed occlusion amblyopia at the second full-time occlusion episode will have two events, with no occlusion amblyopia at the first episode and occlusion amblyopia at the second episode. Because a patient could have multiple events, the logistic model was fitted using generalized estimating equations to account for the correlation of outcomes from the same patient. In addition to the main effects, the logistic regression model also included the interaction effect of amblyopia type and full-time occlusion episode to test for effect of amblyopia type at each full-time occlusion episode and effect of full-time occlusion episode by amblyopia type.
The comparisons involving visual acuity line difference were tested using the Wilcoxon rank-sum test. To test for the association between the number of full-time occlusion episodes to occlusion amblyopia and how many episodes it took for equalization, the Jonckheere-Terpstra test was used to test for increasing or decreasing trends in median number of post-occlusion amblyopia episodes across categories of full-time occlusion episodes to occlusion amblyopia of 1, 2, and 3 or more.
Between 1970 and 2000, 1,541 records of patients with the diagnosis of amblyopia were reviewed. Of those, 597 patients prescribed full-time patching for amblyopia were included and 944 patients were excluded because they were not treated with full-time occlusion (133), had no patching (98), did not have amblyopia (93), were older than 10 years when diagnosed as having amblyopia (176), were lost to follow-up (292), had unavailable records (28), or had amblyopia from causes other than strabismus, anisometropia, or both and with ocular structural abnormalities (124). Two patients did not have adequate follow-up to analyze the episode of when they developed occlusion amblyopia.
Of the total 597 patients treated by full-time patching for amblyopia, 115 were found to meet the criteria for occlusion amblyopia for an overall incidence of 19.3%. The overall frequency of occlusion amblyopia decreased with increasing age; it was highest (37.7%) in patients younger than 12 months and continued to decrease to 3.4% in patients 60 months or older (Table 1). The incidence of occlusion amblyopia by amblyopia type suggests a significant difference between occlusion amblyopia and the types of amblyopia because the frequency was 22.9% in strabismic amblyopia (100 of a total of 437) and 9.4% in the anisometropic/combined amblyopia group (15 of 160) (Table 1).
Table 1: Frequency of Occlusion Amblyopia in the Study Population (N = 597)
Occlusion amblyopia occurred more often after the first episode of patching (Table 2). There was sufficient follow-up for statistical analysis for 595 of the 597 patients. Of these 595 patients, 86 or 14.4% (95% confidence interval: 11.7%, 17.5%) had occlusion amblyopia at the first full-time occlusion episode. Of the 412 who continued to the second full-time occlusion episode, 15 or 3.6% (95% confidence interval: 2.0%, 5.9%) had occlusion amblyopia at the second full-time occlusion episode. Of the 297 who continued to the third full-time occlusion episode, 7 or 2.4% (95% confidence interval: 1.0%, 4.8%) had occlusion amblyopia at the third full-time occlusion episode. Of the 217 who continued to the fourth full-time occlusion episode, 2 or 0.9% (95% confidence interval: 0.1%, 3.3%) had occlusion amblyopia at the fourth episode. Of the 166 who continued to the fifth full-time occlusion episode, 2 or 1.2% (95% confidence interval: 0.2%, 4.3%) had occlusion amblyopia at the fifth episode. Of the 127 who continued to more than five full-time occlusion episodes, 3 or 2.4% (95% confidence interval: 0.5%, 6.8%) had occlusion amblyopia.
Table 2: Comparison of the Frequency of Occlusion Amblyopia With Each Successive Episode of Full-time Occlusion (FTO)
Patients with strabismic amblyopia began full-time patching at a significantly younger age (33 months) than those with combined (47.5 months) or anisometropic (60.5 months) amblyopia. Using a multi-factor regression model comparing incidence of occlusion amblyopia between strabismic and combined types, with age at the beginning of full-time occlusion included in the model, those patients who had their first full-time occlusion at a younger age were more likely to develop occlusion amblyopia, with an odds ratio of occlusion amblyopia for those who started their first full-time occlusion at younger than 36 months of 8.56 (95% confidence interval: 2.73, 26.84; P = .0002) times greater than those who started full-time occlusion at older than 60 months. Children 36 to 60 months had an odds ratio of 2.66 (95% confidence interval: 0.96, 7.37) (P = .059, Table 3). Occlusion amblyopia was also more likely to occur in those with strabismic amblyopia compared to anisometropic or combined amblyopia, although this was only observed at the first episode of full-time occlusion (P = .011; odds ratio of 2.89, 95% confidence interval: 1.28, 6.54). For strabismic amblyopia, occlusion amblyopia was more likely to develop at the earlier episodes of full-time occlusion, with an odds ratio of occlusion amblyopia for the first full-time occlusion episode relative to the fourth or later full-time occlusion episode of 16.21 (95% confidence interval: 3.89, 67.60) (P = .0001, Table 3).
Table 3: Multifactor Logistic Regression for Occlusion Amblyopia
Similarly, the incidence of occlusion amblyopia between those tested by fixation testing (35.0%) and optotype testing (8.0%) was significantly different. However, the median age at initiation of full-time occlusion in those tested by fixation preference (21 months) was significantly less than those with optotype testing (58 months). In a multi-factor logistic regression model comparing incidence of occlusion amblyopia between fixation preference and optotype testing with age at the commencement of full-time occlusion also included in the model, age (P = .0002) was significantly associated with occlusion amblyopia but no significant effect was seen for test method (P = .481; odds ratio of occlusion amblyopia for fixation preference/optotype of 1.31 [95% confidence interval: 0.62, 2.75]) (Table 3).
Of the 115 of 597 patients who developed occlusion amblyopia, visual acuity at time of onset of occlusion amblyopia was obtained by fixation preference testing in 89 patients (31.0% of 287 patients who were tested by fixation preference). The remaining 26 patients (8.4% of the total patients tested by optotypes) were tested with line pictures (15), line Es (6), or line letters (5). Of those tested with optotypes, the median difference in visual acuity between the two eyes when beginning full-time occlusion for patients who developed occlusion amblyopia (5.5 lines) was not significantly different (P = .161) than for those patients who did not develop occlusion amblyopia (4 lines) (Table 4).
Table 4: Comparison of the Interocular Visual Acuity (Line Difference) Between Patients Who Developed Occlusion Amblyopia and Those Who Did Not
Outcome measures were calculated for all patients who had 6 or more months of follow-up after stopping all patching therapy who could complete optotype vision testing. This included 96 of the patients who had developed occlusion amblyopia (83%) and 396 (82%) of those who had not. The final interocular visual acuity line difference was compared between those who developed occlusion amblyopia and those who did not. The median interocular acuity difference was 0 lines in those who developed occlusion amblyopia and 1 line in those who did not. When comparing the frequency distribution of the two groups, the final interocular difference was significantly less (P = .003) in those with a history of occlusion amblyopia compared to those who did not develop occlusion amblyopia (Table 4).
Patients who developed occlusion amblyopia were treated initially with one age-related episode of no occlusion in 74 patients (64%), with occlusion of the originally amblyopic eye in 34 patients (30%) or alternate patching in 7 patients (6%). In general, if visual acuity did not improve after one age-related episode of no patching, then patching of the original amblyopic eye was instituted. Observation alone was successful in reversing occlusion amblyopia or equalizing vision in 39 of the patients (53%) treated initially with observation. Patients whose occlusion amblyopia occurred after the first episode of full-time occlusion appeared to require more episodes of observation or treatment to reverse their occlusion amblyopia or to equalize visual acuity (Table 5).
Table 5: No. of Episodes of Patching Required to Reach Equal Vision Compared to the No. of FTO Episodes After Which Occlusion Amblyopia Was Diagnosed
For patients who developed occlusion amblyopia, final visual acuity in the eye with occlusion amblyopia was compared to final visual acuity in the original amblyopic eye. All patients had visual acuity of either 20/30 or better in the eye with occlusion amblyopia. Sixty-five patients (66%) had equal visual acuity, with 49 of these having visual acuity of 20/20 or better in both eyes; all were at least 20/30. Twenty-six patients (27%) had worse visual acuity in the originally amblyopic eye. Seven patients (7%) had poorer visual acuity in the eye with occlusion amblyopia than in the originally amblyopic eye at the final examination (Table 6). Of these seven patients, all had visual acuity of 20/30 or better in each eye and all but one were within a one-line difference between the two eyes.
Table 6: Data for the Originally Amblyopic Eye and the Eye That Developed Occlusion Amblyopia
Several case reports have documented the occurrence of occlusion amblyopia, often associated with full-time patching.1–5 Eccentric fixation in the occluded eye has also been reported.1,3,5 Some consider the risk of occlusion amblyopia to be a contraindication to full-time patching,4 whereas others believe that occlusion amblyopia is transient and easily reversible and may be a good prognostic sign.1
A few larger reviews have been conducted. In 1969, Goodier8 found 21 cases of occlusion amblyopia among “several hundred” patients treated by total or partial occlusion. These patients were at least 3 years of age and able to complete some form of optotype visual acuity testing. Of these, 14 patients showed a temporary transfer of amblyopia to the originally fixating eye and 7 patients showed a permanent transfer of amblyopia.
In 1982, Assaf9 reviewed 1,904 patients with strabismic amblyopia treated with total continuous, partial, or intermittent occlusion and 115 patients (6%) were found to have a transfer of fixation during treatment. Of these, 25 had a transient switch, 75 had a temporary switch, and 15 had a permanent switch of fixation. Of those 15 patients characterized as having a permanent switch, 10 eventually switched back or alternated fixation with further occlusion. A permanent switch of fixation was found to be more frequent in younger patients and to occur during both total and partial occlusion. One limitation of Assaf’s study and our study, which both include young children, is the difficulty in obtaining optotype visual acuity and the reliance on fixation preference to diagnose amblyopia. Using a switch in fixation preference to indicate occlusion amblyopia had occurred may result in a higher incidence of occlusion amblyopia.
More recently, the Pediatric Eye Disease Investigator Group compared prescribing full-time patching (defined as all hours or all but 1 hour per day) to 6 hours per day of patching for treatment of moderate amblyopia.10 They found a decrease in the visual acuity in the sound eye of two lines or greater from baseline while patching in 3 patients (4%) in the 6-hour group and 9 patients (11%) in the full-time group. Of these 12 patients, 1 patient in the full-time group continued to test one line worse in the patched eye, originally 20/20 to 20/25. In 9 patients, the visual acuity of the sound eye during later follow-up tested the same or better than baseline and the remaining 2 patients were lost to follow-up.
In our study, the overall incidence of occlusion amblyopia with full-time patching was 19%. There were no cases of clinically significant irreversible occlusion amblyopia, although permanent switches of fixation did occur. Incidence appears inversely related to age. It should be noted that our definition of full-time occlusion as strictly applied in this study (24 hours a day or all waking hours) is more rigorous than most.
Our incidence was higher than that mentioned in previous studies. This may be due to the large number of patients in our group who began patching at a young age. When we look only at those who began patching at an age when their initial visual acuity was tested with optotypes, the incidence is only 8%. This is similar to the 11% incidence reported in the full-time occlusion group studied by the Pediatric Eye Disease Investigator Group.
The incidence of occlusion amblyopia appears to be highly correlated to age at the beginning of treatment. Even when factors such as amblyopia type or testing method are considered, age still seems to be the leading factor. The age at which a child is no longer susceptible to the development of occlusion amblyopia is not known. This study shows that occlusion amblyopia can occur at various ages, with one patient developing occlusion amblyopia as late as 6.8 years of age. However, the incidence of occlusion amblyopia decreases significantly with increasing age.
Patients who developed occlusion amblyopia generally developed it early in patching therapy. Seventy-five percent of patients with occlusion amblyopia had onset within the first age-related episode of full-time patching. Although occlusion amblyopia developed quickly, it also resolved quickly. Discontinuation of patching alone reversed occlusion amblyopia or equalized visual acuity in 34% of these patients.
It has been suggested that the development of occlusion amblyopia may result in permanent loss of vision or a change in fixation preference. A few patients did appear to permanently change fixation preference. At the final recorded visual acuity examination, 7 patients continued to have better vision in their previously amblyopic eye, but all had final visual acuity of 20/30 or better in each eye. Overall, the final visual acuity results of those who had an occurrence of occlusion amblyopia were better than those who did not.
Our results suggest that occlusion amblyopia happens quickly, usually during the first episode of patching, and its incidence decreases with age. Because occlusion amblyopia is most common in very young children, it may be beneficial to consider either a shorter follow-up time for these patients or partial occlusion initially. Older patients who have already completed patching for multiple episodes may not need as frequent follow-up visits. When occlusion amblyopia does occur, it is often merely a temporary transfer of fixation that is easily reversed when treated promptly. The development of occlusion amblyopia is associated with excellent final visual acuity outcomes and may simply point to a highly plastic visual system.
This study is limited by its retrospective nature; however, the data were compiled in a consistent manner by physicians and orthoptists over 26 years, using the same treatment protocol or methodology during that time. The study was limited to patients with good compliance; however, this selection bias probably causes an overestimation of the incidence of occlusion amblyopia, if anything, because we assume that less compliant patients are less likely to develop occlusion amblyopia, although that is an untested assumption.
Another limitation is the variety of visual acuity tests used from fixation testing in the youngest children to line pictures, tumbling Es, line numbers, and line letters. This was one reason interocular visual acuity was used over changes in a baseline visual acuity. The reliability of fixation patterns is debated, but is still considered by the authors to be an extremely useful test in observing children with amblyopia too young to complete optotype visual acuity testing. One of the few previous studies of occlusion amblyopia relied exclusively on fixation preference. When both tests were performed at an examination, the optotype results were used.
Occlusion amblyopia can occur with full-time patching but often results in a good outcome for both eyes, better even than in equally compliant patients who never develop occlusion amblyopia.
- Burian HM. Occlusion amblyopia and the development of eccentric fixation in occluded eyes. Am J Ophthalmol. 1966;62:853–856.
- Costenbader FD. Round table discussion. In: Allen JH. Strabismus Ophthalmic Symposium II. St. Louis: C.V. Mosby: 1958:484–485.
- Hardesty HH. Occlusion amblyopia: report of a case. AMA Arch Ophthalmol. 1959;62:314–316 doi:10.1001/archopht.1959.04220020140020 [CrossRef] .
- Simon JW, Parks MM, Price EC. Severe visual loss resulting from occlusion therapy for amblyopia. J Pediatr Ophthalmol Strabismus. 1987;24:244–246.
- von Noorden GK. The etiology and pathogenesis of fixation anomalies in strabismus. Trans Am Ophthalmol Soc. 1969;67:698–751.
- Scott WE, Kutschke PJ, Keech RV, et al. Amblyopia treatment outcomes. J AAPOS. 2005;9:107–111 doi:10.1016/j.jaapos.2004.12.003 [CrossRef] .
- Keech RV, Ottar W, Zhang L. The minimum occlusion trial for the treatment of amblyopia. Ophthalmology. 2002;109:2261–2264 doi:10.1016/S0161-6420(02)01282-4 [CrossRef] .
- Goodier H. A survey of patients showing transfer of amblyopia from the originally deviated eye to the originally fixing eye by means of occlusion therapy. British Orthoptic Journal. 1969;26:103–110.
- Assaf AA. The sensitive period: transfer of fixation after occlusion for strabismic amblyopia. Br J Ophthalmol. 1982;66:64–70 doi:10.1136/bjo.66.1.64 [CrossRef] .
- Holmes JM, Kraker RT, Beck RW, et al. A randomized trial of prescribed patching regimens for treatment of severe amblyopia in children. Ophthalmology. 2003;110:2075–2087 doi:10.1016/j.ophtha.2003.08.001 [CrossRef] .
Frequency of Occlusion Amblyopia in the Study Population (N = 597)a
| Strabismus||43||100 (22.9%)|
| Anisometropia/combined||160||15 (9.4%)|
|Age at 1st full-time occlusion (mo)|
| < 12||69||26 (37.7%)|
| 12 to < 24||103||34 (33.0%)|
| 24 to < 36||99||31 (31.3%)|
| 36 to < 48||93||10 (10.8%)|
| 48 to < 60||88||9 (10.2%)|
|⩾ 60||145||5 (3.4%)|
Comparison of the Frequency of Occlusion Amblyopia With Each Successive Episode of Full-time Occlusion (FTO)a
|Variable||1st FTO||2nd FTO||3rd FTO||⩾ 4th FTO|
|All||595||86 (14.4%)||412||15 (3.6%)||297||7 (2.4%)||217||5 (2.3%)|
| Strabismus||436||79 (18.1%)||278||13 (4.7%)||195||5 (2.6%)||141||2 (1.4%)|
| Anisometropia/combined||159||7 (4.4%)||134||2 (1.5%)||102||2 (2.0%)||76||3 (4.0%)|
|Age at 1st FTO (mo)|
| < 12||69||19 (27.5%)||39||4 (10.3%)||29||3 (10.3%)||21||0 (0.0%)|
| 12 to < 24||103||28 (27.2%)||61||4 (6.6%)||45||1 (2.2%)||36||1 (2.8%)|
| 24 to < 36||99||26 (26.3%)||59||2 (3.4%)||46||1 (2.2%)||32||2 (6.2%)|
| 36 to < 48||92||4 (4.4%)||66||2 (3.0%)||42||1 (2.4%)||28||2 (7.1%)|
| 48 to < 60||87||6 (6.9%)||69||2 (2.9%)||47||0 (0.0%)||37||0 (0%)|
| ⩾ 60||145||2 (2.1%)||118||1 (0.8%)||88||1 (1.1%)||63||0 (0.0%)|
Multifactor Logistic Regression for Occlusion Amblyopiaa
|Variable||Odds Ratio (95% CI)||P|
| 1st FTO (vs > 4th)||16.21 (3.89, 67.60)||.0001|
| 2nd FTO (vs > 4th)||3.66 (0.81, 16.60)||.093|
| 3rd FTO (vs > 4th)||1.90 (0.36, 10.03)||.450|
| 1st FTO (vs > 4th)||1.03 (0.26, 4.09)||.971|
| 2nd FTO (vs > 4th)||0.36 (0.06, 2.19)||.267|
| 3rd FTO (vs > 4th)||0.45 (0.07, 2.81)||.389|
|Strabismus vs anisometropia/combined|
| 1st FTO||2.89 (1.28, 6.54)||.011|
| 2nd FTO||1.87 (0.41, 8.57)||.423|
| 3rd FTO||0.78 (0.14, 4.26)||.775|
| > 4th FTO||0.18 (0.03, 1.15)||.071|
|Age at 1st FTO|
| < 36 (vs > 60)||8.56 (2.73, 26.84)||.0002|
| 36 to < 60 (vs > 60)||2.66 (0.96, 7.37)||.059|
|Fixation (vs optotypes)||1.31 (0.62, 2.75)||.481|
Comparison of the Interocular Visual Acuity (Line Difference) Between Patients Who Developed Occlusion Amblyopia and Those Who Did Not
|Visual Acuity||Occlusion Amblyopia||No Occlusion Amblyopia||Pa|
|Line difference at start of FTO among those tested by optotype||26||284||.161|
| 25th to 75th percentile||3 to 8||2 to 7.5|
| Minimum-maximum||1 to 12||0 to 19|
|Final line difference||96||396||.003|
| 25th to 75th percentile||0 to 1||0 to 2|
| Minimum-maximum||0 to 4||0 to 15|
|Final line difference–frequency distributionb|
| 0||54 (56%)||168 (42%)|
| 1||30 (31%)||124 (31%)|
| 2||7 (7%)||59 (15%)|
| 3||3 (3%)||21 (5%)|
| 4||2 (2%)||10 (3%)|
| ⩾ 5||0 (0%)||14 (4%)|
No. of Episodes of Patching Required to Reach Equal Vision Compared to the No. of FTO Episodes After Which Occlusion Amblyopia Was Diagnoseda
|Episode||FTO Episode After Which Occlusion Amblyopia Was Noted|
|1 (n = 85)||2 (n = 15)||3+ (n = 12)|
|Subsequent episodes to equalization|
| 1||37 (44%)||13 (87%)||8 (67%)|
| 2||37 (44%)||2 (13%)||3 (25%)|
| 3||9 (11%)||0 (0%)||1 (8%)|
| 4+||2 (2%)||0 (0%)||0 (0%)|
|Median episode to equalization||2||1||1|
Data for the Originally Amblyopic Eye and the Eye That Developed Occlusion Amblyopia
|Patient||Final VAa||Amblyopia Type||Occurrence of Occlusion Amblyopia|
|Originally Amblyopic Eye||Occlusion Amblyopic Eye||Age (mo)||FTO Episode|