Interventions for Stimulus Deprivation Amblyopia [Review]

Abstract

Background

Stimulus deprivation amblyopia (SDA) develops due to an obstruction to the clear passage of light, preventing clear formation of an image on the retina for example, cataract, ptosis (droopy eyelid). It is particularly severe and can be resistant to treatment and the visual prognosis is often poor. Stimulus deprivation amblyopia is rare and precise estimates of prevalence difficult to come by; it probably constitutes less than 3% of all cases of amblyopia. In developed countries most patients present under the age of one; in less developed parts of the world presentation is likely to be significantly later than this.

The mainstay of treatment is patching of the better-seeing eye but regimes vary, treatment is difficult to execute and results are often disappointing.

Objectives

The objectives of this review were to evaluate the effectiveness of occlusion treatment for SDA, determine the optimum treatment regime and factors that may affect outcome.

Search strategy

We searched the Cochrane Central Register of Controlled Trials - CENTRAL (which contains the Cochrane Eyes and Vision Group Trials Register) on The Cochrane Library (2006, Issue 1), MEDLINE (1996 to April 2006), EMBASE (1980 to April 2006) and LILACS (Latin American and Caribbean Literature on Health Sciences) (to November 2004). There were no date or language restrictions.

Selection criteria

We aimed to include randomised and quasi-randomised controlled trials of subjects with unilateral SDA defined as worse than 0.2 LogMAR or equivalent. There were no restrictions with respect to age, gender, ethnicity, co-morbidities, medication use, and the number of participants.

Data collection and analysis

Two authors independently assessed study abstracts identified by the electronic searches.

Main results

No trials were identified that met the inclusion criteria.

Authors' conclusions

It is not possible to conclude how effective treatment for SDA is or which treatment regime produces the best results. There is a need for further study in this area.

Background

Introduction

Amblyopia derives from a Greek word meaning bluntness of vision. In the clinical field it is used to denote a reduction in vision in the absence of any ophthalmoscopically detectable retinal anomaly, and in the absence of any disorder of the afferent visual pathways (Duke-Elder 1973).

Amblyopia can be bilateral but is most commonly unilateral. Bilateral amblyopia most typically develops due to a failure to provide optical correction for high degrees of long or short sight whether primary or following removal of cataracts in childhood. Vision tends to be symmetrically reduced and therefore occlusion treatment is rarely indicated. This review only looks at unilateral amblyopia. It is not uncommon for amblyopia to be of mixed aetiology.

Amblyopia is usually classified according to the cause:

• strabismic: as a result of squint (eye misalignment);

• anisometropic: unequal refractive (focussing) error;

• meridional: due to astigmatism (irregular corneal curvature);

• ammetropic: high refractive error in both eyes;

• stimulus deprivation: secondary to an obstacle in the anterior visual pathway.

Where more than one cause exists it will often be described as mixed amblyopia; typically this refers to the most common combination of strabismic and anisometropia amblyopia.

Pathophysiology

The organisation of the adult visual cortex (brain) is determined by early visual experiences (Wiesel 1963). The period within which abnormal visual input can lead to a disruption of the normal pattern of development is called the ‘critical period’ (Hockfield 1998). There are several critical periods, each associated with different visual functions (Harwerth 1990) that probably reflect the development of different parts of the brain. These critical periods do not end abruptly and can be considered as a continuum from extreme sensitivity to almost no sensitivity to external stimuli. Amblyopia finds its roots in these critical periods at young ages when the brain and visual system are immature and connections between neurons are still being formed and stabilised. During the critical period amblyopia is reversible; this is up to approximately eight years of age but varies considerably between patients and depending on the type of amblyopia.

The focus of this review was unilateral stimulus deprivation amblyopia (SDA); interventions for other types of amblyopia are being evaluated in a series of separate Cochrane reviews.

Etiology

Stimulus deprivation amblyopia, also known as amblyopia ex anopsia, refers to the type of amblyopia where loss of vision results from disuse or lack of formation of retinal images, most commonly as a result of one of the following:

• unoperated infantile cataract (opacity of the lens);

• ptosis (droopy lid) (Dray 2002; Gusek 2000);

• hemangioma (blood rich swelling on the lid) (Schulz 1982);

• vitreous hemorrhages (bleeding into the clear gel that fills the eye) (Ferrone 1994), other obstructions in vitreous such as persistent hyperplastic primary vitreous (PHPV);

• aphakia (absence of the natural lens);

• occlusion prescribed to treat amblyopia of the other eye (Awaya 1973; Von Noorden 1973; Von Noorden 1981).

In some cases the eye itself is otherwise healthy; in others there is co-existing pathology such as microphthalmus (small eye), coloboma (incomplete formation of the eye), optic nerve hypolasia (under-developed optic nerve) or retinal abnormality. It can be very difficult to discern to what extent visual loss is due to the amblyopia and what is due to other pathology; co-existing disease will often limit the visual prognosis making treatment harder to manage.

Stimulus deprivation amblyopia resulting from a congenital or infantile unilateral cataract appears to be the most commonly reported type of SDA. The affected eye is subjected to stimulus deprivation secondary to the cataract until the cataract is removed and then continues until optical correction is provided. The aphakic eye may continue to be subjected to anisometropia and aniseikonia (unequal image size) even after optical correction (Enoch 1983). The early insult to the visual system generally makes this type of amblyopia particularly severe and apparently resistant to treatment; the visual prognosis is reported to be poor (Kanski 1994; Taylor 1997).

Epidemiology

The prevalence of amblyopia in the general population is uncertain, with estimates ranging from 1% to 5% (Brown 2000; Hillis 1983). In European children, the prevalence ranges from 1% to 2.5% (Kvarnstrom 2001; Newman 2000). Amblyopia has been reported to account for 29% of unilateral blindness in Copenhagen (Buch 2001) and as much as 8.3% of bilateral blindness in India due to inadequate treatment following childhood cataract surgery (Dandona 2003). Stimulus deprivation amblyopia is a rare type of amblyopia, probably seen in less than 3% of amblyopic patients (Hillis 1983). There are no known age, gender, race, or developing-developed country differences.

Presentation

Routine health checks of babies and toddlers provide an opportunity for the causative signs associated with SDA, such as ptosis or cataract, to be detected. Post-natal screening may be carried out by a variety of personnel, for example pediatricians, nurses, or family doctors although provision of such screening is not universal. Access to health care professionals and services may be limited, especially in rural areas of developing countries.

Stimulus derivation amblyopia itself is not likely to be noticed, but parents may notice the signs associated with the cause of SA such as leucocoria (whitish pupils) with congenital cataracts or the droopy eyelid (ptosis). If poor vision in one eye is established it may develop strabismus or squint (misalignment) and this may lead to referral. In the developed world most patients present for treatment under a year old (Mein 1991) although this is likely to be significantly later where there is more limited healthcare.

Diagnosis

There are four main steps in the diagnosis of SDA.

1. Visual acuity testing. Testing young children is largely reliant on objective observations which are limited by cognition and concentration. Qualitative methods may be used, such as assessing fixation preference in squint; however, quantitative tests such as preferential looking are more precise. Preferential looking tests rely on the observation that infants prefer to look at patterned rather than plain surfaces (Fanz 1958). If the child can discern the striped panel on the card presented, it will look at it. The degree of visual angle subtended by the stripes is known and therefore a Snellen equivalent can be calculated. In older children testing methods are less objective, relying on the child identifying pictures or letter optotypes in Snellen, decimal or LogMAR notation.

2. External and internal eye examination to identify any pathology; some pathology, particularly optic nerve hypoplasia, needs to be carefully looked for in a child. Treatment may be inappropriately and unsuccessfully commenced if such visually limiting pathology remains undetected.

3. Cycloplegic refraction and prescription of glasses if indicated; amblyopia cannot be diagnosed unless any significant refractive error has been corrected.

4. Rechecking visual acuity with any prescribed refractive correction in place; some improvement in visual acuity can be expected with glasses alone and there should be a period of adjustment into glasses before retesting. Traditionally this adjustment period has been 4 to 6 weeks but studies on refractive and strabismic amblyopia show this may often be up to 24 weeks (Moseley 2002).

Definitions of amblyopia vary largely due to the fact that there is little evidence as to what constitutes normal vision on many commonly used tests at different ages. It may be defined by comparing the eyes (inter-ocular difference) or by looking at monocular visual acuity alone. We have elected to define amblyopia as vision of worse than 6/9 on a Snellen based test, or 0.2 LogMAR (or equivalent) in one eye.

Treatment options

The visual loss attributable to SDA can be severe. The aim of treatment is to maximise visual recovery without adversely affecting acuity in the better seeing eye. The rationale for treatment is two-fold: to provide a good second eye should the better-seeing eye ever be visually compromised and to maximise any binocular cooperation between the eyes. Untreated or unsuccessfully treated amblyopia may have an impact in adult life. For individuals with amblyopia the lifetime risk of serious visual impairment, due to loss or damage of the better-seeing eye is estimated to be between 1.2% and 3.3% (Rahi 2002). In addition there are implications for employment prospects and therefore income; the number of jobs barred to individuals with reduced vision increases with the severity of the deficit (Adams 1999).

Stages of treatment

1. Correct the causative factor that is degrading the quality of the visual image, for example removing infantile cataract, raising the ptosis. Most studies and text books agree that in cases of early unilateral form deprivation correction must be undertaken in the first 8 to 12 weeks of life if good visual acuity is to be obtained (Birch 1986; Birch 1988; Gregg 1992; Kanski 1994; McCulloch 1994; Taylor 1997)

2. Prescribe any necessary glasses to make the quality of visual stimulation received by the child's amblyopic eye as good as possible. Where cataract surgery has been performed intraocular implants or contact lenses, or both may be used.

3. Occlusion therapy: occlusion forces use of the amblyopic eye, stimulating the formation of functional connections in the brain (Boothe 2000).

Occlusion regimens

Protocols and practices vary considerably; duration of occlusion therapy varies from full-time to as short as one hour. Factors affecting the amount prescribed include the level of vision deficit, the age of the child and the likely waiting time to the next appointment. Follow-up is recommended at intervals of one week per year of age during periods of aggressive patching (Simon 1987). Occlusion can be stopped when visual acuity becomes equal in the two eyes or if no progress has been made after three months of good compliance with occlusion (Pratt-Johnson 2001). It has been recommended that children in this situation are monitored up to the age of visual maturity (approximately 7 years of age) to ensure that amblyopia does not recur. Some periods of maintenance occlusion may be required during that time (Mein 1991).

The following have been used as additions to occlusion therapy but generally appear to no longer be used:

1. CAM visual stimulator: uses rotating high-contrast square wave gratings to stimulate the amblyopic eye.

2. Pleoptics: employs after-images to encourage foveal fixation and normal projection in the amblyopic eye

Types of occlusion

Atropine penalisation and optical penalisation (use of lenses to reduce the acuity) are other forms of occlusion that encourage use of the amblyopic eye by diminishing visual form. These treatments for amblyopia are being evaluated in another Cochrane review currently underway.

This review examined the role of total occlusion to form and light as an intervention for SDA. Total occlusion, also known as conventional occlusion, is usually achieved by means of an opaque, adhesive patch on the better seeing eye. Less commonly, occlusive contact lenses are employed. As mentioned previously bilateral SDA is rare and not usually treated with occlusion therapy and therefore was not considered in this review.

Measuring outcomes

Qualitative methods for assessing the presence of amblyopia in preverbal children are based on the observation of fixation patterns. These methods require highly trained examiners and are often unreliable (Wright 1986; Zipf 1976). In order to quantify amblyopia a measure of visual acuity must be obtained; final visual acuity assessed using an age-appropriate test (Fulton 1978; Sebris 1987) is the most commonly used outcome from treatment. Tests vary in the use of optotypes (picture, letter, symbol) and may be with or without crowding; crowded visual acuity tests are harder to perform but are more sensitive to amblyopia than uncrowded tests.

Developmental changes in young children complicate the evaluation of actual change in acuity from pre to post treatment. Alternative methods of measuring change have been suggested in an attempt to overcome this (Schmidt 1994; Stewart 2003) but we aimed to compare post treatment visual acuity values, defining restoration of normal visual acuity as better than or equal to 6/9.5 on Snellen or 0.2 LogMAR (or equivalent).

Factors affecting outcome

Compliance with therapy is critical for successful treatment but can often be very difficult to achieve. Young children can become distressed by being restricted to reduced visual acuity and the discomfort of wearing an adhesive patch. It has been suggested that, if possible, compliance should be monitored in order to more effectively measure response to treatment. Devices to objectively measure compliance have been developed (Awan 2005; Stewart 2005) but are not yet in common use and clinicians generally depend on parental reports. Other factors that are thought to affect the success of treatment are the duration of the visual deprivation and age at the onset of therapy (Maurer 1989): the earlier onset, longer duration and the later treatment is commenced the worse the visual prognosis.

Harm from occlusion therapy

Potential adverse effects from occlusion therapy include inducing amblyopia in the occluded eye, skin allergies, infections or corneal abrasions (or both) from contact lens wear, intractable diplopia (double vision), and psychological effects such as distress.

Rationale for a systematic review

The reported success of treatment for SDA varies. There are studies reporting good levels of vision following early treatment (Gregg 1992; McCulloch 1994) but there is no standardisation and poor agreement among experts as to the optimum amount of occlusion needed to achieve good visual outcomes. Commencing occlusion therapy in infants with very poor vision can be harrowing for the parents and stressful for the child and realistic treatment goals are often poorly defined. It is necessary to establish what degree of improvement can realistically be expected and the most effective occlusion regime(s) for achieving this.

Objectives

The principal objective was to evaluate the effectiveness of occlusion therapy in an attempt to establish realistic treatment outcomes. Where data was available, we also examined evidence for any dose/response effect and assessed the impact of the duration, severity and causative factor on the size and direction of the treatment effect.

Criteria for Considering Studies for this Review

Types of studies

This review aimed to include randomised and quasi-randomised trials.

As no randomised controlled trials were found, other relevant studies already known to the authors or identified in the searches have been described in order to comment on current practice. It is important to note that these were not systematically searched for so do not represent a systematic summary of existing evidence.

Methods for future updates

If any randomised or quasi-randomised trials are identified in future updates of this review we will follow the criteria for inclusion and methods for analysis as set out below.

Types of participants

• unilateral stimulus deprivation amblyopia (SDA) defined as best corrected visual acuity of worse than 6/9.5 Snellen (or equivalent) after any treatment for the causative factor has been undertaken and refractive error corrected. (Other co-existing amblyogenic factors will be reported).

• no restrictions with respect to age, gender, ethnicity, co-morbidity, medication use, and the number of participants.

Co-existing ocular pathology that may limit the visual outcome, for example coloboma, optic nerve hypoplasia or retinal dystrophy, will be documented. Studies including such participants will be subjected to a subgroup analysis.

Where methods for determining the age at onset, and the duration of the deprivation are reported these will be described. If there is any uncertainty we will attempt to contact the authors for clarification.

It is likely that there will be variation in the time allowed for adjustment into glasses. Where possible the adjustment period will be reported, and the implications discussed.

Types of interventions

The following interventions are of interest:

• total occlusion by adhesive patch such as conventional occlusion;

• total occlusion by occlusive contact lens;

• pleoptic treatment;

• CAM visual stimulation.

We planned to examine the following comparisons:

1. total occlusion versus no occlusion;

2. any means of total occlusion compared to another;

3. any total occlusion plus pleoptic treatment versus total occlusion alone;

4. any total occlusion plus CAM visual stimulator versus total occlusion alone;

5. full-time occlusion (> 6 hours / day) versus part-time occlusion (≤ 6 hours / day).

Types of outcome measures

Primary outcomes

The primary outcome for this review is best-corrected visual acuity of the amblyopic eye, on an age-appropriate test, six months from cessation of occlusion.

Although not directly equivalent, Snellen data was converted into a LogMAR equivalent for ease of interpretation and analysis.

Outcomes will be dichotomised:

1. Normal = 0.2 LogMAR (6/9.5 Snellen or equivalent) or better.

2. Residual deficit = worse than 0.2 LogMAR.

   Where possible mean values will be reported.

Secondary outcomes

The secondary outcomes for this review are:

1. Visual acuity in the amblyopic eye at seven years of age or older.

2. The proportion of the amblyopia deficit corrected (Stewart 2003).

3. Any measure of stereoacuity (3 dimensional vision).

Cost data

Any trials detailing the comparative costs of treatment methods will be described.

Adverse effects

We will summarise the reported adverse effects related to treatment. Any reports of the following will be described:

Severe - occlusion amblyopia; contact lens related problems, for example infection, corneal abrasions; adverse psychological effects, for example distress, treatment ceased due to poor compliance or failure to attend; intractable diplopia.

Minor: allergy to patches.

Quality of life measures

When available, data on quality of life measures will be described.

Follow up

A minimum of six months post treatment follow up is necessary for inclusion and analysis; eligible studies with less follow up will be described.

Search Strategy for Identification of Studies

Electronic searches

We searched the Cochrane Central Register of Controlled Trials - CENTRAL (which contains the Cochrane Eyes and Vision Group Trials Register) on The Cochrane Library, MEDLINE, EMBASE and LILACS (Latin American and Caribbean Literature on Health Sciences). There were no date or language restrictions.

The following strategies were used.

• CENTRAL: 2006, issue 1

• #1 AMBLYOPIA

• #2 PUPIL DISORDERS

• #3 CATARACT

• #4 BLEPHAROPTOSIS

• #5 VITREOUS HEMORRHAGE

• #6 APHAKIA

• #7 (amblyopia or (ex next anopsia))

• #8 cataract*

• #9 (stimul* or vision* or visual or optical) and (deprivat*)

• #10 (capillary and (hemangioma* or haemangioma*)

• #11 (media near opacity)

• #12 (vitreous and (haemorrhage* or hemorrhage*)

• #13 (cornea* and (scar* or opac* or degenerate*)

• #14 (ptosis or blepharoptosis)

• #15 (#1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12 or #13 or #14)

• #16 patch*

• #17 (optical or visual)

• #18 (stimul* or penalisation)

• #19 (occlus* or pleoptic*)

• #20 (therap* or treatment* or lens* or complete* or partial*)

• #21 #16 or #17 or #18 or #19 or #20

• #22 #15 and #21

• MEDLINE: 1966 to April 2006

• #1 explode “Amblyopia-” / all SUBHEADINGS in MIME,MJME

• #2 explode “Pupil-Disorders” / all SUBHEADINGS in MIME,MJME

• #3 explode “Cataract-” / all SUBHEADINGS in MIME,MJME

• #4 explode “Blepharoptosis-” / all SUBHEADINGS in MIME,MJME

• #5 explode “Hemangioma-Capillary” / all SUBHEADINGS in MIME,MJME

• #6 explode “Vitreous-Hemorrhage” / all SUBHEADINGS in MIME,MJME

• #7 explode “Aphakia-” / all SUBHEADINGS in MIME,MJME

• #8 amblyopia or ex anopsia

• #9 (stimul* or vision* or visual or optical) near3 (deprivat*)

• #10 cataract*

• #11 (capillary) near3 (h?emangioma*)

• #12 (media) near3 (opacity)

• #13 (vitreous) near3 (h?emorrhage*)

• #14 (corneal) near3 (opacity or scar* or degenerat*)

• #15 ptosis

• # 16 or #1- #15

• #17 patch*

• #18 (optical or visual) near3 (stimuli* or penali?ation)

• #19 (occlus* or pleoptic*) near3 (therap* or treatment* or lens* or complete or partial)

• #20 (vi?ual*) near3 (stimulat*)

• #21 #17 or #18 or #19 or #20

• #22 #16 and #21 (not restricted to randomised controlled trials (RCTs))

• EMBASE: 1980 to April 2006

• #1 exp AMBLYOPIA/

• #2 exp Pupil Disease/

• #3 exp CATARACT/

• #4 exp Ptosis/

• #5 exp CAPILLARY HEMANGIOMA/

• #6 exp Vitreous Hemorrhage/

• #7 exp APHAKIA/

• #8 #1 or #2 or #3 or #4 or #5 or #6 or #7

• #9 ((stimul$ or vision$ or visual$ or optical$) adj3 deprivat$).mp.

• #10 cataract$.mp.

• #11 (media adj3 opacity).mp.

• #12 (vitreous adj3 (hemorrhage$ or haemorrhage$)).mp.

• #13 (corneal adj3 (opacity or scar$ or degenerat$)).mp.

• #14 (ptosis or blepharoptosis).mp.

• #15 amblyopiamp or exp anopsia/

• #16 (amblyopia or ex-anopsia).mp.

• #17 #8 or #9 or #10 or #11 or #12 or #13 or #14 or #15 or #16

• #19 patch$.mp.

• #20 ((optical$ or visual$ or vizual$) adj3 (stimul$ or penalisation or penalization)).mp.

• #21 ((occlus$ or pleoptic$) adj3 (therap$ or treatment$ or lens$ complete$ or partial$)).

• #22 #19 or #20 or #21

• #23 #17 and #22

• To limit the search to RCTs this search strategy was combined with the following:

• #1 Randomized Controlled Trial/

• #2 Clinical Trial/

• #3 exp Randomization/

• #4 Double Blind Procedure/

• #5 Single Blind Procedure/

• #6 Controlled Study/

• #7 (control or compar$) adj3 (group$ or subject$ or patient$)).ab,ti.

• #8 exp ANIMAL/

• #9 Human/

• #10 #8 not #9

• #11 #1 or #2 or #3 or #4 or #5 or #6 or #7

• #12 #11 not #10

• LILACS was searched to November 2004 using the following strategy.

• #1 amblyopi$ or aphakia or anopsia

• #2 ((stimul$ or vis$ or viz$ or optic$) and deprivat$)

• #3 child$ or adolesc$ or juvenile$ or minor$ or school$ or kindergarten$ or pre-school$ or nurser$ or infan$ or baby or babies

• #4 #1 or #2

• #5 #3 and #4

Other sources

No manual searches were undertaken for this review but will be carried out if possible in future updates.

Methods of the Review

Assessment of search results

Two review authors independently assessed the titles and abstracts of all reports identified by the electronic searches as per the ‘Criteria for considering studies for this review.’ The authors were unmasked to the report authors, institutions and trial results during this assessment.

The abstracts were classified as (a) definitely include, (b) unsure and (c) definitely exclude. Full copies of those classified as (a) definitely include and (b) unsure were obtained and re-assessed. The studies were then classified as (1) included, (2) awaiting clarification and (3) excluded. The concordance between authors was documented; a third review author resolved any disagreements. Authors of studies classified as (2) awaiting assessment were contacted for further clarification. Details of studies identified by both authors as (3) excluded were documented in the relevant section of the review.

The following methods will be adopted for future updates of this review.

Assessment of methodological quality

Two review authors will assess the sources of systematic bias in trials according to methods set out in section 6 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2005). The following parameters will be considered:

• quality of allocation concealment (selection bias);

• method of randomisation;

• completeness of follow up such as attrition bias - how many participants were lost to follow up, how they were accounted for, whether follow up rates for groups were similar;

• whether all participants were analysed as randomised. If studies report that an intention-to-treat analysis (ITT) was performed, we will assess whether both a) participants where no outcome was collected, and b) those who only received some or none of their allotted treatment were included. We will only interpret a true ITT analysis to have been undertaken if both these criteria have been fulfilled.

• detection bias: whether assessment of outcome was concealed and if so, how adequately.

Each of the parameters will be graded as (A) Adequate or Yes, (B) Unclear or Not Reported, and (C) Inadequate or No. Agreement between authors will be documented; a third author will resolve any disagreement. Masking of participants and care providers is not feasible in these trials and hence will not be used as a measure of quality. For trials categorised as (B) Unclear or Not Reported, the authors will be contacted for additional information. If the authors do not respond, the authors will assign a grade to the trial based on the available information.

Assessment of study characteristics

In addition to the parameters described above, data will be extracted on:

1. Participants - numbers, age at onset and intervention, duration of stimulus deprivation, cause of stimulus deprivation, starting visual acuity (and test type), refractive correction.

2. Intervention - method of occlusion, regime, CAM, pleoptics.

3. Outcomes - test(s) used, length of follow up, if, when and how compliance assessed.

Data collection

Two review authors will independently extract the data for the primary and secondary outcomes onto paper data collection forms developed by the Cochrane Eyes and Vision Group. Discrepancies will be resolved by discussion. Primary investigators will be contacted for missing data. One author will enter all data into RevMan 4.2. A second author independently will re-enter the data, using the double data-entry facility to check for inaccuracies.

Data synthesis

Data analysis will follow the guidelines in section 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2005). For dichotomous outcomes we will calculate a summary odds ratio for rarer outcomes or risk ratio for more frequent outcomes. Weighted mean difference will be calculated for continuous outcomes such as if any trials are identified that have measured vision using LogMAR tests throughout the study. We will test for statistical heterogeneity (chi-square test and using the I-square value) and if no statistical heterogeneity is detected and if there is no clinical heterogeneity within the trials we will combine the results in a meta-analysis using a fixed-effect model. If statistical heterogeneity is present in the absence of clinical heterogeneity we will consider whether or not it is appropriate to compute a summary measure. It is likely that a fixed-effect model will be used if the number of trials is three or less. In case of substantial statistical or clinical heterogeneity (I-square value greater than say 50%) we will not combine study results but present a tabulated or narrative summary.

Subgroup analysis

If sufficient numbers of trials are available and they are stratified prior to randomisation the following subgroups will be explored:

• participants without any co-existing ocular pathology (that might be expected to limit visual prognosis) will be analysed separately from those with other pathology.

• participants with stimulus deprivation amblyopia associated with a unilateral congenital cataract/ compared to stimulus deprivation amblyopia associated with any other unilateral etiology.

Sensitivity analysis

Sensitivity analyses will be conducted, if appropriate, to determine the size and direction of effect when excluding the following:

• outcomes measured on uncrowded vision tests;

• studies where any parameter has been graded ‘C’ or ‘No’.

If appropriate, analyses will also be carried out excluding unpublished studies or industry-funded studies.

Description of Studies

The electronic searches identified 799 abstracts and titles of which seven appeared to be randomised controlled trials of interventions for amblyopia. Three of these (Clarke 2003; Holmes 2003; Repka 2003) evaluated conventional occlusion therapy but did not include patients with stimulus deprivation amblyopia (SDA). Four studies (Keith 1980; Mehdorn 1981; Nyman 1983; Tytla 1981) were trials of the CAM visual stimulator but after reading the full text of the studies and contacting the authors where necessary it became apparent that only one trial (Nyman 1983) had included participants with SDA and the data relevant to this review were no longer available. All seven trials were therefore excluded (see the ‘characteristics of excluded studies’ table).

Full text copies of a further 25 references were obtained because the initial search information was insufficient to establish whether the studies were eligible for inclusion or not. This was either because the title only was available, the abstract was unclear or the abstract or study was written in a language other than English. After further perusal or translation all studies were found to be ineligible and excluded. Reasons for exclusion have been documented in the ‘characteristics of excluded studies’ table.

Methodological Quality of Included Studies

As no randomised controlled trials were included, none were assessed for methodological quality.

Results

None of the studies identified in the searches were eligible for inclusion highlighting a significant gap in the existing evidence for the treatment of stimulus deprivation amblyopia. In order to provide the reader with some insight into the basis for current practice, some of the non-randomised studies identified incidentally in the searches and others already known to the authors are commented on in the discussion.

Discussion

Treatment for the more common forms of amblyopia has recently, and for the first time, been subject to more rigorous scrutiny in high quality randomised controlled trials. These studies have helped to clarify at what level amblyopia treatment works and have provided useful information about which occlusion regimes may work most effectively (Clarke 2003; Holmes 2003; Repka 2003). However, along with most other non-randomised studies on amblyopia, stimulus deprivation amblyopia (SDA) is nearly always specifically excluded. This is because SDA is generally accepted not only to be more severe and therefore more resistant to treatment (Kanski 1994; Taylor 1997), but may also (based on animal studies) have a different pathophysiological mechanism from the other types (Mitchell 2002). Treatment for SDA is confounded by many factors. The frequent presence of co-existing pathology, the young age of the patient and the limitations of clinical tests make it very difficult not only to quantify the degree of visual deficit but to establish how much is attributable to amblyopia and whether or not it is responding to treatment. The age of the patient and the severity of visual loss can also result in poor compliance with treatment and significant stress and distress for the parents and the patient. However even allowing for these real-life obstacles to good visual recovery there is a dearth of evidence as what outcomes might realistically be expected.

Current evidence of treatment is largely derived from non-randomised studies of SDA caused by unilateral congenital cataract. A brief overview of some of these data is summarised below.

Occlusion type

As might be expected, the majority of studies we came across described the use of total or conventional occlusion for the treatment of SDA. Although it is not without disadvantages in terms of discomfort, it is relatively easy to control the dosage of treatment and is without the more complex side effects of occlusive contact lenses. One randomised trial (Nyman 1983) looked at the additional effect of CAM stimulation compared to total occlusion alone but the data for SDA could not be isolated from data for the other types of amblyopia. This treatment, though prevalent in the 1980s has largely disappeared from current practice, possibly due to the time implications for patient and clinician against the lack of evidence of long-term benefit.

Occlusion regimes and treatment outcomes

Current practice generally appears to favour aggressive patching in early life based on the knowledge that the visual system is much more sensitive to change at this age. Mayer 1989 reports a negative correlation between the number of hours patched and interocular difference in acuity. Intensive or aggressive patching generally means a minimum of six hours per day and may be as much as 100% of waking hours. Birch 1988 reports 53% achieving 20/80 (6/24) or better with this treatment. Lundvall 2002 found 20% attained 0.1(6/7.5) or better and Drummond 1989 reported 43% achieved better than 20/50 (6/9). Robb 1987 found 46% achieved at least 20/70 (6/18). Although by no means comprehensive, this brief summary highlights the variable ‘success’ rate and also the variety of ways in which results can be categorised. This and other variations in study methodologies makes it impossible to meaningfully compare results between these studies.

Less intense occlusion regimes, while also being easier to execute have been advocated as allowing more binocular interaction and therefore a better chance of developing stereoacuity. Brown 1999 reported good visual and binocular results with one hour per day per month of age for the first six months of life.

Compliance

Many, if not all papers on occlusion treatment for SDA highlight the necessity of good compliance if there is to be any chance of achieving a satisfactory outcome. Simultaneously most clinicians are aware that compliance in treating SDA is extremely difficult to achieve. While it may not be too surprising that a treatment that visually compromises a child by means of an adhesive patch is not easy to deliver, justification of such a treatment must carefully consider any potential harm alongside evidence of benefit. In a culture where justifying intervention is increasingly required the current absence of clear evidence of effectiveness in this area is concerning. Compliance also effects interpretation of the dose-response treatment effect. Some studies on refractive and strabismic amblyopia have used objective methods to monitor how much occlusion is actually worn (Awan 2005; Loudon 2002; Stewart 2005). These show that the prescribed amount of occlusion is not always achieved but also that the lower doses of occlusion can be as effective as more intense occlusion; data which will help progress towards establishing the optimum dosage required for maximum treatment benefit. To our knowledge, no such studies have yet been conducted on people with SDA.

Authors' Conclusions

Implications for practice

It is not possible to draw reliable conclusions from the available data as the study designs either do not compare treatment strategies or are subject to significant bias in the selection of participants for particular treatments. In addition, the variation between studies as to how treatment is delivered and outcomes are categorised prevents the comparison or combination of results.

The general trend in practice (based on the proportion of papers we found reporting this treatment) appears to be in favour of more intensive occlusion therapy for better visual outcomes but this has not been thoroughly tested and is linked to problems with compliance. There is some evidence that less intense treatment may have favourable results. The difficulties associated with treatment, the demand on resources and the potential impact on the patient need to be considered carefully against the current absence of real evidence of treatment benefit in this condition. It is currently difficult to objectively advise parents or formulate evidence based guidelines for the management of this condition and it remains very uncertain what it is realistic to expect from treatment for stimulus deprivation amblyopia and how best to achieve it.

Implications for research

There is a clear and pressing need for higher quality studies in this area of clinical practice. While occlusion therapy currently remains the mainstay of treatment and withdrawing it for a time in the context of a randomised controlled trial may be deemed unethical, it must also be considered whether continuing an intense and potentially traumatic treatment without a clearly defined end point or clear evidence of effectiveness might not also be deemed the same.

Unsuccessful treatment ultimately results in the same outcome as no treatment - blindness or partial sight in one eye (or in the case of trauma to the good eye, even both eyes). However the exposure to treatment carries with it the potential for harm and future studies of treatment for this condition must seriously consider and accurately measure any potential harm whether physical, emotional or psychological as well as reporting treatment benefit.

Specific questions that need to be addressed in prospective, randomised studies (with appropriate pre-randomisation stratification for any subgroup analyses) are:

• what level of vision can be realistically achieved;

• how effective is full-time versus part-time occlusion;

• what duration of treatment is necessary to achieve optimum benefit;

• what are the potential harms from treatment;

what factors are associated with satisfactory and unsatisfactory outcomes.

Plain Language Summary.

Patching treatment for a type of lazy eye caused by obstructed vision in early childhood

The term ‘lazy eye’ is used to describe amblyopia or poor vision which cannot be corrected by wearing glasses. There are various causes. Stimulus deprivation amblyopia (SDA), the type examined in this review, develops due to vision being obscured in early childhood by conditions such as cataract (cloudy lens) or ptosis (droopy eyelid); it is generally accepted to be the hardest type of amblyopia to treat. The reported prevalence of amblyopia varies from 1% to 5%; SDA probably constitutes less than 3% of all cases of amblyopia. The presenting sign of SDA is usually the obstruction to vision, for example the ptosis or cataract, rather than the reduced vision itself. Health professionals may detect the abnormality or parents may notice signs such as the whitish pupil associated with congenital cataract. Most patients present under the age of one. Amblyopia is diagnosed after the causative factor has been treated and any necessary glasses are being worn. The level of vision taken to be below normal varies: we elected to use vision below 0.2 LogMAR or equivalent although typically the level of loss in SDA is much more severe. The aim of treatment is to maximise visual recovery without adversely affecting the better-seeing eye. The rationale is to provide a good second eye should the better eye ever lose vision and to maximise cooperation between the eyes. Untreated or unsuccessfully treated amblyopia may affect choice of employment in adult life. Treatment is only effective in early childhood. Patching the better-seeing eye is the mainstay of treatment but prescribed duration per day varies from all waking hours to as little as one hour. The reported success of treatment also varies. Some studies report good outcomes following early treatment but there is no standardisation and poor agreement as to the optimum treatment regime. Patching treatment can be harrowing for parents and stressful for the child. The aim of the review was to examine existing evidence to help establish realistic treatment outcomes and the most effective treatment regime(s). We searched for randomised controlled trials examining the effectiveness of patching or comparing treatment strategies but did not find any that fulfilled our inclusion criteria. There remains a pressing need for better quality evidence of treatment effectiveness in this condition.

Characteristics of Excluded Studies

Study	Reason for exclusion
Arruga 1966	Review, not a clinical trial.
Clarke 2003	RCT but stimulus deprivation amblyopia not included.
Cramer 1966	Not a randomised controlled trial.
Cuppers 1967	Retrospective case-control study*.
Fletcher 1969a	RCT but stimulus deprivation amblyopia not included.
Fletcher 1969b	Retrospective chart review.
Flynn 1967	Retrospective chart review.
Flynn 1968	Retrospective study.
Funghini 1973	Non-comparative study.
Holmes 2003	RCT but stimulus deprivation amblyopia not included.
Iacobucci 1977	Review article.
Keith 1980	Tril of CAM vision stimulator; stimulus deprivation amblyopia not included.
Kuming 1982	Before-after study with only 2 participants with stimulus deprivation amblyopia.
Lang 1965	Case-series, non-comparative*
Lennerstrand 1983	No participants with stimulus deprivation amblyopia.
Mackensen 1965	Case-series, non-comparative*.
Malik 1970	Cohort study.
Mehdorn 1981	Trial of CAM vision stimulator RCT; stimulus deprivation amblyopia not included.
Nyman 1983	Trial of CAM vision stimulator; data on stimulus deprivation amblyopia included but could not be extrapolated and no longer available.
Pistelka 1973	Non-comparative study.
Priegnitz 1965	Non-comparative study*.
Repka 2003	RCT but stimulus deprivation amblyopia not included.
Schor 1983	Did not include participants with stimulus deprivation amblyopia.
Shroff 1983	Non-comparative study.
Stojcevska 1975	Non-comparative study*.
Tomlinson 1973	Non-comparative study.
Tommila 1969	Non-comparative study.
Tommila 1974	Review article.
Tytla 1981	Trial of CAM vision stimulator; stimulus deprivation amblyopia not included.
Veronneau 1974	Used historical controls.
Widder 1967	Non-comparative study*.
Zang 1988	Non-comparative study.

^*- full-text articles of these studies published in non-English languages were reviewed and are noted in this table