Predicting Onset and Progression of Myopia: Axial Length Percentile Curves

Refractive error assessments, with or without the aid of cycloplegia, remain the cornerstone of refractive error management and their value in assessing and managing myopia remains undisputed. However, they are not without issues.

…we believe, these percentile charts have utility in screening for myopia and estimating the probability of myopia and future risk of progression, which is of great value when educating both patients and parents

Figure 1.
1. The percentile curves are spaced in increments ranging from the fifth to 95th percentiles. Being placed in a higher percentile indicates that a given eye has a longer eye length relative to the reference population, for example, the 70th percentile indicates that 29 out of 100 children have a longer eye than the individual. In the chart, for a child aged 10 years, both right and left eyes are placed in the higher percentiles, indicating a higher probability for myopia (current) as well as a higher risk of future progression.
2. A model based on axial length, axial length/corneal curvature, age and gender is used to estimate spherical equivalent at an adult range and is provided as a range within which the refractive error may lie.
3. The risk for myopia based on the percentile and estimation of spherical equivalent is summarised, and in the current example is ‘high’.

In young children with active accommodation, non-cycloplegic assessments risk misdiagnosis or a more myopic estimation of refractive error.1 Additionally, accuracy/repeatability of non-cycloplegic and subjective refractive error assessments is variable and ranges ± 0.50D between techniques.²

Although cycloplegic assessments are more accurate, and are considered to be the gold standard for refractive error assessments, eye care professionals in many countries do not have access to mydriatics and cycloplegics, and therefore rely on noncycloplegic measurements. Furthermore, cycloplegic assessments also come with limitations. They require more chair/clinic time as well as additional assessments in the clinic to rule out potential adverse events, which necessitate more resources. The use of mydriatics/cycloplegia result in poor vision for a period that limits the individual’s activities and may also impose a burden on carers/parents. Additionally, the risk of adverse effects with certain mydriatic and cycloplegic agents necessitates close monitoring.

Against this background, the introduction of optical biometers that use partial coherence interferometry and swept source techniques, enable rapid, accurate and noninvasive measurements of axial length.³ Importantly, axial length measurement is accurate to within a few microns that translate to +/-0.12D and additionally, many of these devices also enable measurements of corneal curvature (CR) and/or refractive error measurements.

THE VALUE OF OPTICAL BIOMETERS

Increasing availability of optical biometers begs the question: How valuable are axial length (AL) and axial length/corneal curvature (AL/CR) measurements when screening and managing myopia?

Although an increased AL is the underlying basis for myopia, AL/CR is considered a more robust measure of the refractive status as it also considers the contribution of the cornea.⁴ Indeed, we found that a model that incorporates AL, AL/CR, age and gender had good sensitivity/specificity (87.4/88.2) for detection of myopia in an Asian cohort.⁵ The aforementioned value is slightly greater when compared with models that consider non-cycloplegic refraction alone (sensitivity/specificity of 85.7/87.4), or non-cycloplegic refractive error in combination with visual acuity (85.1/88.6).⁶ This indicates that AL and AL/CR values have potential benefit when screening high risk populations for myopia. For example, they may be useful in school or community-based screenings where rapid assessment with a biometer will help identify those at risk. The at-risk individuals may then be sent for a more comprehensive examination.

Also useful are AL and AL/CR percentile curves in estimating the risk of myopia as well as indicating future progression. Using a large data set of approximately 14,127 Asian children, ranging in age from four to 18 years, and approximately 7,204 European eyes ranging from four to 24 years-of-age, we developed growth curves (percentile curves) to illustrate the risk. These percentile curves are employed in the Oculus Myopia Master.

A FAMILIAR TOOL

Growth charts are familiar to everyone and are commonly used to illustrate the distribution of weight and height in infants and young children. They provide an easy and visual reference, with each curve on the chart representing a percentile of axial length, across ages, from the reference population.

Figure 2. The Oculus Myopia Master.

The percentile curves can be used to rank the position of an eye’s axial length by indicating the percent of the reference population the child’s eye would equal or exceed. The percentile charts can be used to estimate risk (both current and future) in both myopic and non-myopic eyes (Figure 1) and differ from calculators such as the BHVI Myopia calculator (bhvi.org/myopia-calculator-resources) which estimates risk of progression in myopic eyes only.

Impressively, we believe, these percentile charts have utility in screening for myopia and estimating the probability of myopia and future risk of progression, which is of great value when educating both patients and parents, and managing the condition.

Professor Padmaja Sankaridurg is Head, Myopia Program and Head, Intellectual Property at the Brien Holden Vision Institute, an IMI advisory board member, and Conjoint Professor at the School of Optometry and Vision Science, UNSW, Sydney, Australia. 

Associate Professor Thomas Naduvilath works with the Brien Holden Vision Institute, NSW School of Optometry and Vision Sciences. His work focusses on statistical modelling of ocular events, ocular comfort and myopia progression. He is currently co-investigator of myopia multi-ethnic study. 

References 

1. Sankaridurg P, He X, Naduvilath T, et al. Comparison of noncycloplegic and cycloplegic autorefraction in categorizing refractive error data in children. Acta ophthalmologica. Nov 2017;95(7):e633-e640. doi: 10.1111/aos.13569. 
2. Rosenfield M, Chiu NN. Repeatability of subjective and objective refraction. Optometry and vision science : official publication of the American Academy of Optometry. Aug 1995;72(8):577-579. 
3. Song JS, Yoon DY, Hyon JY, Jeon HS. Comparison of Ocular Biometry and Refractive Outcomes Using IOL Master 500, IOL Master 700, and Lenstar LS900. Korean journal of ophthalmology : KJO. Apr 2020;34(2):126-132. doi: 10.3341/kjo.2019.0102. 
4. Grosvenor T, Scott R. Role of the axial length/corneal radius ratio in determining the refractive state of the eye. Optometry and vision science : official publication of the American Academy of Optometry. Sep 1994;71(9):573- 579. doi: 10.1097/00006324-199409000-00005. 
5. He X, Sankaridurg P, Naduvilath T, et al. Normative data and percentile curves for axial length and axial length/ corneal curvature in Chinese children and adolescents aged 4-18 years. The British journal of ophthalmology. Sep 16 2021. doi: 10.1136/bjophthalmol-2021-319431. 
6. Lin S, Ma Y, He X, Zhu J, Zou H. Using Decision Curve Analysis to Evaluate Common Strategies for Myopia Screening in School-Aged Children. Ophthalmic epidemiology. Aug 2019;26(4):286-294. doi: 10.1080/09286586.2019.1616774.