Multifocal spectacles in childhood myopia: Are treatment effects maintained? A systematic review and meta-analysis

Methodological quality of the review: Medium confidence

Author: Dinesh Kaphle, David A. Atchison, Katrina L. Schmid

Region: China, Hong Kong, USA, Japan, Canada, South Korea

Sector: Myopia

Subsector: Myopia treatment with multifocal spectacles

Equity focus: No

Study population: Children

Type of programme: School based

Review type: Other review

Quantitative synthesis method: Systematic review and meta-analysis

Qualitative synthesis method: Not applicable

Background: A recent meta-analysis estimated that nearly half of the world’s population will be myopic by 2050, of which 10 per cent will have high myopia. In parts of Asia, the prevalence of myopia has increased dramatically in just one or two generations. Effectiveness of the myopia control intervention is usually expressed as the absolute treatment effect. This is the mean difference in change in spherical equivalent refraction (SER in diopters) or axial length (in mm) between intervention and control groups, as opposed to the percentage (relative) treatment effect, which is one hundred times the difference divided by the change in the control group. Percentage treatment effect may be an intuitive way to discuss myopia treatment options with parents, but it has limitations. Most myopia control studies present the treatment effect between baseline and the end of the study period, but some studies provide myopia progression at more than two time points. As the treatment effect depends on the duration of the study, direct comparisons cannot be made between studies with different time periods. Therefore, this review evaluated the treatment effects of multifocal spectacles on myopia control over both six- and 12-month intervals so that comparison of the treatment effects could be made across studies using the same treatment duration. Understanding the treatment effects over time might contribute to the determination of ways to maintain the efficacy of optical interventions.

Objectives: The aim of this study was to evaluate the impact of duration on the treatment effect of multifocal spectacle lenses used to inhibit myopia progression in children.

Main findings: Nine randomised control trials involving 1,701 children aged 8 to 13 years were included in the meta-analysis. The length of follow-up varied from 12 to 36 months: two studies were 12 months, one study was 18 months, two studies were 24 months, one was 30 months, and three studies were 36 months. Treatment effects, that is, differences in spherical equivalent refraction between intervention and nonintervention groups, were analysed over both 6- and 12-month intervals. As treatment duration increased, effectiveness reduced. In six-month intervals, treatment effects were 0.07 D (95 % CI 0.02; 0.13), 0.03 D (95% CI -0.02, 0.08), and 0.02 D (95% CI -0.05, 0.11) for baseline to 6, 6-12, and 12-18 months, respectively. For 12-month intervals, treatment effects were 0.21 D (95% CI 0.12, 0.29), 0.11 D (95% CI 0.03, 0.19), and 0.12 D (95% CI -0.01, 0.25) for baseline to 12, 12-24, and 24-36 months, respectively. Even during the second six months of wear, the ability of multifocal spectacle lenses to inhibit myopia progression was reduced.

Authors note that it is not appropriate to extrapolate the treatment effect observed in the first six months or 12 months to estimate the likely future benefit of treatment.

Methodology:

Inclusion criteria consisted of: 1) randomised controlled trials; 2) multifocal spectacles as the intervention and single-vision lens spectacles as the control; 3) primary outcome was change in cycloplegic spherical equivalent refraction, that is, myopia progression, with 95% confidence interval (CI) or standard deviation at six- or 12-month intervals; 4) participants were children with myopia; 5) treatment duration of at least 12 months.

Literature searches were performed on PubMed, Embase and the Web of Science for articles published up to 14 February 2019. Articles performed in “humans” and published in English language were used as filters. The reference lists of original reports and review articles retrieved during the search were reviewed for additional studies not yet included in the computerised databases. References lists of included studies were scanned. The following information was extracted by the authors: surname of the first author, publication year, study country, type of multifocal lenses, age and sex of the study population, sample size, proportion lost to follow-up, length of follow-up, and myopia progression with standard deviation at six- or 12-month intervals. The risks of bias associated with the randomisation (selection bias), blinding or masking (performance bias), and loss to follow-up (attrition bias) were assessed using the Jadad score. The random-effects method was used for meta-analysis as it incorporates an assumption that the different studies are estimating different yet related intervention effects. Statistical heterogeneity among articles was assessed with the I² statistic, with I² ≥50% considered to indicate high heterogeneity.

Applicability/external validity: Regarding external validity, the authors reported that this meta-analysis review is likely to have publication bias, as only published studies were included. This could overestimate the treatment effect of intervention as unsuccessful trials may not be published. Also, myopia progression over six- and 12-month intervals was primarily described based on change in SER. It is important to consider the change in axial length as an indicator of myopia progression. Another limitation is that the number of studies in the third interval for both six- and 12-month interval analyses were fewer than that for the first and second intervals.

Geographic focus: Not discussed.

Summary of quality assessment:

There is medium confidence in the conclusions about the effects of this study. Although authors conducted thorough searches of the literature, language bias was not avoided. In addition, screening of studies for inclusion and data extraction was conducted by one reviewer only.