consideration of testing results before adjusting therapy is warranted.

A recent article has also confirmed that the visual acuity may vary from test strategy to test strategy. The ATSHOTV protocol overestimated the visual acuity relative to the E-ETDRS protocol (0.68 lines for amblyopic eyes; 0.25 lines for fellow eyes) [25].

Fixation preference testing has long been the clinical method of choice (in fact the only method in widespread clinical use) for determining amblyopia in children unable to perform a quantitative acuity on an eye chart. The examiner determines the preference for fixation in a strabismic patient simply by determining the eye being used. For the orthotropic patient, a strabismus is created with a 10or 12-prism diopters vertical prism and the assessment of fixation preference is again made. If the patient alternated or at least could hold with the lesspreferred eye through a blink or a pursuit movement, no amblyopia was felt present. Two recent reports using the same testing protocol have found that the test is much less reliable than we have thought. These research groups tested children 30 to less than 72 months with fixation preference testing and optotype acuity. Fixation preference testing identified only 15% of preschool children who had an IOD of two lines or more on visual acuity testing and 25% of those with an IOD of three lines or more [26]. There were an insu cient number of children with strabismus to comment on that subgroup.

In the Multiethnic Pediatric Eye Disease Study (MEPEDS), the authors reported sensitivity of fixation preference testing for amblyopia among children with anisometropia was 20% (9/44), although specificity was 94% (102/109). Among strabismic children, sensitivity was 69% (9/13; worse in children 30–47 than 48–72 months old), and specificity was 79% (70/89) [27].

Hakim found that 75% of strabismic children had positive test results by fixation preference testing, but only 13% had an IOD of two lines or more [28]. The obvious, albeit controversial confusion, is that fixation preference testing misses most cases of amblyopia when used in a screening setting. In addition, the use of fixation preference testing in a clinical setting for managing a patient with strabismus would likely lead to substantial overtreatment.

10.1.5Natural History

Limited natural history data are available for amblyopia as nearly all patients diagnosed are prescribed some therapy. Although compliance is quite variable, most children receive some intervention. Some authors have

suggested a tendency to spontaneous improvement of the visual acuity deficit associated with amblyopia [29, 30]. Alternatively, another research group found that patients who did not comply with treatment deteriorated over time [31]. It is safe to comment that we do not know enough about the natural history of this common condition.

Summary for the Clinician

■Current estimates of the prevalence of amblyopia among preschool aged children in the Unites States range from 0.8 to 2.8%, with the highest rate found among Hispanic Americans. Most cases are associated at least in part with refractive error.

■Fixation preference testing for amblyopia is unreliable for the detection of amblyopia. It also appears to not be su ciently reliable to guide amblyopia therapy in many children.

■Care is needed when interpreting sequential measurements of visual acuity when made with di erent instruments or testing paradigms.

10.2 Amblyopia Management

Best practice for management of amblyopia had been based on clinician consensus [1]. However, no randomized trial had ever been done comparing no treatment to any amblyopia treatment. During the last 5 years, a large number of clinical trials assessing methods of amblyopia treatment have allowed the incorporation of evidencebased information into the practice of amblyopia care based on the earlier guidelines.

10.2.1Refractive Correction

The value of an accurate refraction can not be underestimated in the management of amblyopia. These data are essential for both the diagnosis of amblyopia and the subsequent optimum treatment of the amblyopia. For security of the amblyopia diagnosis, the presence of an anisometropia helps substantiate the presence of amblyopia. The refractive error requires a measurement obtained under adequate cycloplegia, usually 1% cyclopentolate or similar cycloplegic. Many clinicians instill a topical anesthetic before the cycloplegic agent to prolong the retention of the cycloplegic drug in the tear film.

Prescribed glasses for ametropia are not controversial. The prescription for an esotropia patient should be full plus power [32]. Even if this power slightly blurs distance vision, it will not have a deleterious e ect at the child’s

10 usual working distance. For the microstrabismic or orthotropic child, under correcting the hypermetropia symmetrically by up to 1.50 diopters avoids the problem of distance blur and does not seem to detract from the treatment outcome. For the exotropic patient, the anisometropia and any myopia need to be corrected. High hypermetropia should be partially corrected.

What has been controversial among clinicians is what to do (and when) once the eyeglasses prescription is written and spectacles obtained. Some clinicians have routinely started patching at the same time, while others have waited a variable amount of time. Recent research has provided some guidance on this clinical decision, specifically the value of glasses alone in the management of amblyopia. In the United Kingdom, Stewart et al found a mean improvement of 2.4 lines in 65 children 3–8 years of age were treated with spectacles, taking an average of 14 weeks to reach best visual acuity [33]. Surprisingly, improvement was noted among both anisometropic and strabismic patients. These authors have termed this e ect refractive adaptation, although that term is potentially confusing since the refraction does not actually adapt. Rather the improvement represents the remediation of the amblyopia by optical correction alone. In a larger recent prospective study investigators in North America enrolled 84 children 3 to <7 years old with untreated anisometropic amblyopia ranging from 20/40 to 20/250 [34]. Optimal refractive correction was provided in accordance with consensus guidelines similar to those above. VA was measured with the new spectacle correction at baseline and at 5-week intervals until VA stabilized or amblyopia resolved.VA improved with optical correction alone by ≥2 lines in 77% of the patients and remarkably resolved in 27% [34]. Although the study was designed and powered for children with anisometropia, strabismic and combined strabismic–anisometropic patients were enrolled in a parallel pilot study following the same protocol to determine if such patients could respond to spectacle correction alone [35]. Twelve patients with previously untreated strabismic amblyopia were prescribed spectacles and examined at 5-week intervals until visual acuity was not improved from the prior visit. Amblyopic eye acuity improved by ≥2 lines from spectacle-corrected baseline acuity in 9 (75%), resolving in three. Mean change from baseline to maximum improvement was 2.2

± 1.8 lines. Improvement continued for up to 25 weeks. Data on the ocular alignment after instituting the glasses were not available. Improvement in the visual acuity of

amblyopic strabismic patients was not expected to occur so often so PEDIG has launched an adequately powered prospective study of the impact of spectacle correction alone to explore this result.

10.2.2Occlusion by Patching

The beneficial e ect of occlusion with an adhesive patch in the management of amblyopia has long been considered obvious. Some randomized-controlled treatment trials have compared treatments, without an untreated control, led to criticism that the improvements experienced were due to age or learning e ects or possibly the benefits of spectacles alone as noted earlier [36]. To address that issue, PEDIG conducted a RCT comparing occlusion to spectacles only. Before enrollment, the patients wore glasses until their vision stabilized between two consecutive visits. They were then randomized to continue spectacles alone compared with 2h of daily patching. Improvement in VA of the amblyopic eye from baseline to 5 weeks averaged 1.1 lines in the patching group and 0.5 lines in the control group (P = 0.006), and improvement from baseline to best measured VA at any visit averaged 2.2 lines in the patching group and 1.3 lines in the control group (P < 0.001) [37]. Thus, occlusion was better but surprisingly there was continuing benefit of the spectacles alone, reinforcing how important this aspect of therapy must be.

The dosage of occlusion therapy prescribed has historically ranged widely, from a few minutes to all waking hours per day. Some clinicians have prescribed fewer hours for fear of damaging the binocular visual system. In the initial PEDIG trial, comparing atropine to patching, both treatments were found to be equally e ective [38]. Subgroup analysis of di ering dosages from 6 h daily to full time (all waking hours less one daily) found no advantage of prescribing more hours [39]. This led us to design two studies directed at exploring occlusion dosage. In the first trial, we compared 2 with 6 h daily for the initial treatment of moderate amblyopia, 20/40–20/80, for a period of 4 months [40]. Visual acuity in the amblyopic eye improved a similar amount in both groups. The improvement in the amblyopic eye from baseline to 4 months averaged 2.40 lines in each group (P = 0.98). The 4-month visual acuity was ≥20/30 and/or improved from baseline by ≥3 lines in 62% in each group (P = 1.00). We did not follow and treat these patients after 4 months so we do not know if a di erence might develop. In the second trial of patching dosage, we compared 6 with full time or all waking hours less 1 h for severe amblyopia, 20/100–20/400 [41]. VA in the amblyopic eye improved to a similar extent in both groups. The improvement in

the amblyopic eye acuity from the baseline to 17 weeks averaged 4.8 lines in the 6-h group and 4.7 lines in the full-time group (P = 0.45). However, 75% of patients in both groups were 20/40 or worse after therapy. There is a natural concern about amblyopia therapy, particularly with higher dosages, causing loss of vision in the sound eye. The sound eye lost two or more lines in 4% of the 6-h group and in 11% of the full-time group. Nearly all patients returned to their baseline level with follow-up, typically by just stopping all patching.

These patching dosage data show that for initial treatment of amblyopia due to strabismus, anisometropia or both combined, beginning with the lower dosage of occlusion does not lessen the chance of success and may make the treatment more feasible. However, only about one in four patients with moderate amblyopia was 20/25 or better and one in four children with severe amblyopia was 20/32 or better.

These studies have taught much about initial patching therapy, but they have left substantial uncertainty about what to do for those children who are not completely corrected. Some clinicians have misinterpreted the results and have recommended stopping therapy when the visual acuity ceases to improve with these prescribed doses. What needs to be explored is whether an increased dose or a change in treatment approach will allow more complete correction. At present, clinicians and parents will have to make that judgment without the results of a RCT to guide the choice. Logically, some period of more intense therapy should be administered before discontinuing treatment.

10.2.3Pharmacological Treatment with Atropine

To find an e ective, yet easy to administer, treatment of amblyopia has been a goal pursued by clinicians treating amblyopia in response to the complaints and di culties associated with occlusion therapy. This pursuit has led to many failed treatments that were launched with great fanfare, but ultimate abandonment.

For more than a century, clinicians have used pharmacological penalization of the sound eye to make the child use the amblyopic eye and thereby improve the visual acuity of that eye. Most clinicians typically used this treatment for patching failures or noncompliance. Case series reported e ectiveness, but the common belief was that this was an inferior treatment. The largest prospective study was completed in 2002, comparing once daily atropine to patching 6 or more hours per day for moderate amblyopia 20/30–20/100 [38]. Visual acuity improved in

both groups: 2.84 lines in the atropine group and 3.16 lines in the patching group. The patching group did get better faster, but by 6 months, the di erence of 0.034 was clinically inconsequential. Both treatments were well tolerated, although the atropine was easier to administer based on parental questionnaires.

These children were followed in the study for an additional 18 months to describe prescribed treatment and stability of the improvement. Treatment was determined by the investigator [42]. Remarkably, and at odds with clinical wisdom, nearly 90% received some treatment during this period. Eighty percent received the same treatment and 25% received the alternate treatment (some patients received both). At 2 years, visual acuity in the amblyopic eye improved a mean of 3.6 lines in the atropine group and 3.7 lines in the patching group. This difference in visual acuity between treatment groups was small: 0.01 logMAR (95% confidence interval, −0.02 to 0.04). Thus, the relative equivalence of the techniques and the persistence of the treatment benefit were rea rmed. Stereoacuity outcomes were similar suggesting no untoward relative e ect of either of the two treatments.

One concern regarding amblyopia therapy is the potential for inducing or worsening a strabismus. In addition, most authors have suggested treating amblyopia before undertaking strabismus surgery. This study evaluated the chance of inducing a strabismus, but also the chance of improving a strabismus with amblyopia treatment. Of the 161 patients with no strabismus, similar proportions initially assigned to the patching and atropine groups developed new strabismus by 2 years (18 vs. 16%, P < 0.84) [43]. Of the new cases of strabismus, only two patients in the patching group and three patients in the atropine group developed a deviation that was greater than 8D. Perhaps surprisingly, of the 105 patients with strabismus greater than 8D at enrollment, 13% of those in the patching group and 16% of those in the atropine group improved to orthotropia without strabismus surgery. These data show that strabismus may develop or resolve with amblyopia therapy in about equal proportions.

The dosage of atropine in the original PEDIG trial was once daily. This design was consistent with the desire to maximize the likelihood of finding benefit if there was one. While that study was underway, the benefit of less frequent administration was suggested by Simons and coworkers [44]. They reported reasonable improvement from less frequent administration. This was plausible since the duration of cycloplegia was often more than 1 day. This finding led PEDIG to develop a clinical trial, which compared daily atropine to weekend atropine.

The atropine dosage treatment trial included 168 children younger than 7 years with amblyopia in the range of 20/40–20/80 associated with strabismus, anisometropia or both. They were randomized to either daily or week-

10 end atropine [45]. The improvement of the amblyopic eye from baseline to 4 months averaged 2.3 lines in each group. The visual acuity of the amblyopic eye at study completion was either (1) at least 20/25 or (2) better than or equal to the sound eye in 39 children (47%) in the daily group and 45 children (53%) in the weekend group. The visual acuity of the sound eye at the end of follow-up was reduced by two lines in one patient in each group. Stereoacuity outcomes were similar in the two groups.

Patients who were not cured continued on the randomized treatment beyond the 4-month outcome exam. They improved an average of 0.8 additional lines (0.7 lines among the 22 daily group patients and 0.8 lines among the 31 weekend group patients).

At the time of study completion, 39 (47%) of the patients in the daily group and 45 (53%) in the weekend group had an amblyopic eye acuity that was either (1) 20/25 or better or (2) the same or better than the sound eye acuity, provided that the sound eye acuity had not decreased from enrollment. The mean amblyopic eye acuity at study completion was 0.23 logMAR in the daily group and 0.21 logMAR in the weekend group (approximately 20/32). The mean sound eye visual acuity at enrollment was 0.05 logMAR (approximately 20/25), with 81% of the sound eyes having acuity of 20/25 or better.

Among patients who improved two or more lines from baseline during the study, 30% of patients achieved their best acuity at 5 weeks, 50% at 4 months, 7% at 6 months, 10% at 8 months and 3% at 10 months. These results were similar in the two atropine treatment groups. Thus, a 4-month treatment period with atropine will treat most patients but is not su cient to complete treatment for all. Thus, treatment should be continued until there is good evidence that a plateau in improvement has been achieved.

There is a chance of visual impairment of the sound eye so care needs to be taken. In this study 1% of sound eyes lost two or more lines of acuity at last follow up. As expected, light sensitivity was common, reported by 16% of children. Facial flushing and fever, a more worrisome side e ect, was reported by 1% of the children.

Summarizing, weekend atropine for moderate amblyopia is e ective in improving visual acuity. The amount of improvement was comparable with that seen with 4 months of 2 or 6 h of daily patching [40]. Parents need to realize that most children will need at least 4 months of treatment irrespective of which therapy and dosage. Twice weekly atropine is fairly unobtrusive for preschool

children, should easily be incorporated into a child’s daily activities, and is likely to be attractive to a large proportion of parents. However, as with patching if the visual acuity improvement is not complete increasing the dosage or changing to an alternative therapy should be considered. The e ectiveness of such a treatment remains to be proven.

10.2.4Pharmacological Therapy Combined with a Plano Lens

Investigators have long looked for ways to intensify their treatments, implicitly recognizing that the prescribed therapy did not always have the desired e ect. For atropine penalization of the sound eye, it has been long noted that adding optical penalization, by removing all hypermetropic correction from the sound eye, would add optical blur at distance to complement the cycloplegic blur provided at near. A retrospective report included 42 children (mean age, 4.7 years) treated with daily atropine and a plano lens for the sound eye [46]. Important caveats were that eligible patients had failed patching treatment and had at least 1.75 D of sound eye hypermetropia. Surprisingly, they found a mean improvement in amblyopic eye visual acuity from 20/113 to 20/37 after 10 weeks of treatment with atropine and a plano lens to the sound eye. This was a remarkable achievement. However, Morrison and colleagues cautioned that this treatment resulted in a case of severe treatment-related amblyopia in the sound eye when parental noncompliance occurs [47].

To explore the value of this “augmented atropine approach,” PEDIG randomized 180 children with moderate amblyopia (visual acuities of 20/40–20/100) to weekend atropine use augmented by a plano lens or weekend atropine use alone [48]. At 18 weeks, amblyopic eye improvement averaged 2.8 lines in the group that received atropine plus a plano lens and 2.4 lines in the group that received atropine alone (mean di erence between groups adjusted for baseline acuity, 0.3 line; 95% confidence interval, −0.2–0.8 line). Amblyopic eye visual acuity was 20/25 or better in 24 patients (29%) in the group that received atropine only and 35 patients (40%) in the group that received atropine plus a plano lens (P = 0.03). However, more patients in the group that received atropine plus a plano lens had reduced sound eye visual acuity at 18 weeks; fortunately, there were no cases of persistent reverse amblyopia. The important conclusion is that in spite of intuition, augmentation of weekend atropine use with a plano lens does not substantially improve amblyopic eye visual acuity when compared with weekend atropine use alone.