Purpose: In evidence-based medicine, well-conducted meta-analyses provide the highest quality
of scientific evidence. Donovan et al published a meta-analysis of refractive progression
among myopes; 1 however, axial length may be a preferred metric to refractive
error for assessing myopia as
(i) it is significantly associated with retinal
(ii) its measurement is more repeatable (in relative terms) and
it can be obtained without cycloplegia.
We conducted a meta-analysis to model
axial elongation in myopic children.
Methods: A search using PubMed and Embase, based on the following search terms ["myopi* AND
((axial AND (length OR elongation)) OR biometr*) AND (progression OR shift OR longitudinal)
NOT (animal OR surgery OR surgical OR adult)], yielded 269 results. Data from 63 eligible
studies was available after eliminating reviews, studies without progression data
specifically for myopes in the age range of 7 to 18 years and papers with redundant
information or other sundry issues, and adding studies cross-referenced in the papers
surveyed. To account for varying ages of sample populations and different periods
of follow-up, instantaneous axial elongation rate was modeled using the integral of
the exponential decay function based on the following equation:
= (eÎ»t' - eÎ»t) * Î±/Î», where AEt,t'
is measured axial elongation between ages t and t' and Î± and
Î» are fitted constants. A procedure was developed to score race in different
studies with ethnicity, country of residence and proportional representation in studies
taken into account.
Results: Graph A of the figure plots adjusted residuals after modeling instantaneous rate of
axial elongation (mm/yr) by age and race (see caption for details). 79% of measured
values fall within ± 0.10 mm of fitted values. Graph B plots the model (R 2
= 0.82). Asian children show about 44% greater rate of elongation than non-Asians
across age despite independence of fit of the decay functions. A 9yo Asian child is
predicted to have instantaneous progression of 0.41 mm/yr compared with 0.28 mm/yr
for a 9yo white child. These rates of axial elongation are generally compatible with
refractive progression curves of Donovan et al 1, although differences between
Asian and non-Asian children are greater than expected.
Conclusion: To our knowledge, this is the first meta-analysis of axial elongation in myopic children.
Axial length measurement is expected to become an increasing part of clinical practice
as myopia control becomes more widely accepted. Our model provides a useful referent
for practitioners when assessing myopic progression in children and may assist in
setting efficacy criteria for myopia control products. Future work will identify important
covariates beyond race and age, the change in the axial length/refractive error ratio
and progression rates during various forms of myopia control treatment.
Donovan et al, Optom Vis Sci 2012; 89: 27.