Effectiveness of early in comparison to late(r) treatment in children with amblyopia or its risk factors: a systematic review

Methodological quality of the review: High confidence

Author: Schmucker C, Kleijnen J, Grosselfinger R, Riemsma R, Antes G, Lange S, Lagrèze W.

Region: Europe and United States of America (USA)

Sector: Amblyopia, refractive error

Sub-sector: Amblyogenic risk factors, treatment, strabismus

Equity focus: Children or juveniles (up to the age of 18)

Review type: Effectiveness review

Quantitative synthesis method: Narrative analysis

Qualitative synthesis methods: Not applicable

Background

Amblyopia is defined as two line or greater difference in visual acuity (VA) between the two eyes with no demonstrable abnormality of the anterior visual pathways. Risk factors associated with amblyopia include uncorrected refractive errors, in particular anisometropia and strabismus. Treatment of amblyopia consists of correcting refractive errors, and if acuity is not improved, patching treatment by occlusion of the better eye. Although treatment for amblyopia is considered to be the most effective in early childhood, there is some evidence from observational studies that improvements in acuity can occur in some affected adults.

Research objectives

To evaluate the effectiveness of early in comparison to late(r) treatment in children with (1) amblyopia or (2) its risk factors, such as refractive errors and strabismus.

Main findings

The authors included five studies (11 publications) in this systematic review including data from Europe, including United Kingdom (UK) and United States of America (USA). Of these, three were randomized studies (one direct comparison between early treatment and late(r) treatment within one study, and one indirect comparison between two studies). The primary intervention was treatment of amblyopia. The remaining studies (two direct comparisons within one study) evaluated treatment in children with amblyogenic risk factors such as hyperopia and esotropia. These two studies applied a non-randomized study design.

In regards to treatment of amblyopia: one comparison suggested that amblyopia treatment in pre-school children is more effective than treatment later in life. However, a subgroup analysis of children who have never received any treatment indicated that patching may also have an effect after the “sensitive phase”. The second comparison showed that a delay in treatment until the age of five did not seem to influence effectiveness.

Treatment of amblyogenic risk factors: two comparison showed that hyperopia treatment and strabismus surgery (outcome: remaining amblyopia) is more effective under two years of age than later in life. However, the studies showed methodological weakness, for instance a high loss to follow-up, unmasked outcome assessment, limiting the validity of their findings. The gathered literature did not provide data evaluating quality of life or school performance.

The authors noted that, uncertainties remain in regards to the age at which treatment for amblyopia or its risk factors is most effective.

Besides methodological weakness of the literature, study designs made it difficult to address the research question. As such, the authors suggested that future research work should provide rigorously controlled prospective data on whether treatment in preschool years confers any benefit over treatment at school entry or ever later. One approach would be to introduce screening programmes at different time points in different regions. In addition, the authors noted that the impact of amblyopia treatment on family life and the child should form part of the outcome assessment of such studies.

Methodology

The authors included studies that focused on children or juveniles (up the age of 18) from the general population with amblyopia or its risk factors; randomized and non-randomized trials were considered for direct comparisons of early and late(r) treatment. Non-randomised studies had to consider at least three confounding factors.

The authors conducted a search on eight bibliographic databases including Medline, Embase, CINAHL, PsycINFO and CENTRAL from inception until 2007. No restriction to a specific language was applied. The search included a number of terms associated with organic eye disorders, diagnostic tests accuracy and the effectiveness of screening programmes. The searches were also supplemented by manual searching the bibliographies of included studies. Authors of studies were also contacted for additional information.

All stages of study selection, data extraction and quality assessment of included studies were conducted by two reviewers independently. The methodological quality was assessed using the CRD guidelines, taking into consideration the following parameters: methods of randomization and allocation concealment, masking of outcome measurement, sample size planning, group comparability, confounding factors and transparency of patient flow. Each parameter was graded as adequate/yes, unclear/not reported or inadequate/no.

The authors conducted a narrative synthesis of findings rather than a meta-analysis or sensitivity analysis due to the heterogeneity of the included studies.

Applicability/external validity

Overall, authors noted that the literature showed that uncertainties remained about the age at which treatment of amblyopia or two of its risk factors (hyperopia and esotropia) was most effective.

Geographic focus

The authors did not restrict the search to specific income settings. Studies included originated from the UK, USA and other parts of Europe which was not specified by the authors. Nevertheless, findings may not be applicable to low- and middle-income settings due to the scarce of evidence/research in this field.

Quality assessment

There is high confidence in the conclusions about the effects of this study. Authors conducted a thorough search of the literature so we can be confident that relevant studies were not omitted, avoiding risk of bias. Appropriate methods were used in terms of study selection, data extraction and quality assessment of included studies, minimizing the risk of bias. The authors appropriately analysed findings of included studies and acknowledged that due to the heterogeneity, a meta-analysis and sensitivity analysis were not appropriate. Due to the methodological weaknesses of the gathered literature authors did not draw strong policy conclusions.