Bayesian meta-analysis of myopia control with multifocal lenses

Authors: Varnas S, Gu X, Metcalfe A.

Geographical coverage: No stated restriction to studies included. Having reviewed 10 included studies, one was conducted in China.   

Sector: Biomedical

Sub-sector: Treatment

Equity focus: Children

Study population: Children with myopia aged between 6 and 15 years.

Review type: Other review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background: Multifocal spectacles and contact lenses represent the interventions to reduce the progression of myopia with the largest number of RCTs carried out.

Objectives: To provide reliable guidelines for the mean percentage efficacy together with the 95% credibility interval in slowing down progression of myopia by a specific intervention (namely, multi-focal lenses) over defined time periods, derived from a substantial number of randomised controlled clinical trials with consistent outcomes.

Main findings

The overall results of the study indicate that the use of progressive addition spectacle lenses (PALs) consistently slows down the progression of myopia at each follow-up, for a period of up to three years, in a statistically significant manner. The magnitude of this effect remains steady for the initial 12 months but appears to diminish thereafter.

This review’s meta-analyses took into account 10 randomised controlled trials (RCTs) involving 1,662 children to test the effectiveness of Progressive Addition Lenses (PALs). For comparison, the authors analysed nine RCTs involving 982 children who were trialling Soft Multifocal Contact Lenses (MFCLs) in a separate model. The study found that the highest efficacy in slowing down the progression of the scaled sphere equivalent refraction was achieved after a 12-month follow-up, with an average reduction in progression of 28% and a 95% credibility interval between 21% and 35%. In comparison, the 95% credibility interval for the average efficacy of soft MFCLs at the 12-month follow-up was 21% to 37%. Both multifocal spectacles and contact lenses moderately slow down the progression of myopia, compared to Single-Vision Spectacle Lenses (SVLs), in the first 12 months after intervention. However, the relative efficacy of PALs appears to diminish after the first 12 months.

One limitation of the meta-analysis is that it provides estimates of the efficacy of the treatment in the cohort of children with a mean age of 10.0 years old and an average refraction of -2.6 D. At least one of the PAL trials found that the percentage efficacy in a subgroup of younger children (6 to 9 years old) in the first 12 months was almost twice as high as in the whole group (57% vs. 30%). Consequently, the mean efficacy percentages are of limited use in predicting the efficacy for each individual patient. They just establish a benchmark against which other interventions can be assessed.

Since both types of multifocal lenses analysed appear to have very similar mean efficacies and are likely to have the same or very similar mechanism of efficacy, the authors recommend to eyecare practitioners the inclusion of PALs with higher addition powers than have been used routinely to date.

Methodology:

Authors selected relevant clinical trials of PALs based on the following criteria: (1) prospective studies with randomised parallel controls wearing SVLs; (2) included children aged between 6 to15 years old; (3) assessed progressive addition lenses with a fixed addition power ≥1.50 D; (4) outcomes measured were cumulative myopia progression (change in SER) and, optionally, axial elongation (change in AL) from baseline at different visits.

The authors do not provide any details on methods used to identify the included RCTs. Data for the meta-analysis was extracted independently by two reviewers and cross-checked. The following was compiled from all studies: authors, publication year, study design, age of participants, sample size, length of follow-up, near addition of the PALs used, and reported outcomes (both primary and secondary, if available) at each follow-up.

The quality of the 10 trials of PALs was assessed using Jadad scores, and the trials with an overall score of 4 or above were classified as high-quality trials. Those with Jadad scores lower than 4 were classified as “equivocal” and were excluded from the meta-analyses of the high-quality subset of the data.

A Bayesian hierarchical model was used for the analysis of the change in spherical equivalent refraction (SER), when wearing multifocal treatment lenses, relative to the change in SER, when wearing SVLs. A similar analysis was performed for changes in axial length (AL). The treatment effect is modelled as the ratio of the change in refraction, or axial length, using multifocal lenses to the corresponding change using SVLs. Therefore, a treatment effect of 1.0 represents no benefit and a treatment effect of 0.6, for example, represents a 40% reduction in progression.

Applicability/external validity:

Study authors do not explore the applicability or external validity of review findings.

Geographic focus: The authors do not discuss the generalisability of findings specifically to LMICs.

Summary of quality assessment:

Despite the use of rigorous methods to critically evaluate studies, there is a lack of information about the selection process for inclusion (for instance, whether language restrictions were applied, which databases were searched, or if unpublished sources were considered). There’s also no indication that two independent reviewers were involved in the study selection. While the data analysis methods were generally robust, the apparent limitations in the search strategy lead us to have low confidence in the review’s findings.