IMC2019 Part 2 – Axial Lengths and Myopia Calculators


Research Abstract Summary

Welcome to Part 2 of 4 fabulous updates from the 2019 biannual International Myopia Conference! Read on for more about predicting axial length growth, refractive growth and translating statistics into real world progression. Click on the headings to link to the published papers relevant to the conference presentation (where available). Pictured above are notable myopia researchers Mark Bullimore (left) and Noel Brennan (right) who seem to like my husband Paul almost as much as I do! Mark and Noel's latest research on understanding myopia control efficacy is described below.

How well does the BHVI Myopia Calculator predict real world myopia progression?

Kathryn Saunders and team compared their large scale NICER (Northern Ireland Childhood Errors of Refraction) study data with the Myopia Calculator predictions. The vast majority of these children in the NICER cohort are Caucasian, and the follow up period was at least 3 years and up to 6 years.

In younger children aged 9-10 years, 16% progressed within limits of the myopia calculator’s 95% confidence intervals. 58% ended up less myopic than the predictor. None ended up more myopic and 26% were classified as ‘erratic’ - ‘failing to follow a consistent pattern with respect to the Calculator’s confidence intervals’. On average these children, on long term follow up, were 1D less myopic than predicted by the Calculator.

In children aged 12-13 years in the NICER study, 56% ended up within limits of the Calculator; 37% ended up less myopic; 7% ended up more myopic and none were ‘erratic’. These children were on average 0.75D less myopic than predicted by the Calculator on long term follow up.

Older patients (more than 12 years) and higher baseline SER were more accurately predicted by the Calculator.

The authors concluded that when using the Calculator in Caucasian and younger children (less than 12 years), this overestimation should be considered and the Calculator used with caution. This isn’t unexpected given the meta-analysis on which the Calculator is based features data from studies of only 1-2 years duration. That being said it’s the first example of such a Calculator, and big cheers to the BHVI for doing it - with more data and longer studies included, where available, hopefully the predictability can be tested further against a variety of populations and further validated.

Can research give us growth charts for axial length?

This would be a game changer for both tracking myopia progression, but even more importantly, predicting the future myope. Ian Flitcroft, Sara McCullough, James Loughman and Kathryn Saunders have done a whole bunch of complex maths to analyse previously published refraction and axial length data in younger (age 6-9) and older (age 15) cohorts compared to the well known NICER study undertaken on children aged 6-19 years. They found that their algorithm successfully predicted refraction six years later - persistent teenage hyperopes were in the bottom 20% and myopes in the top 80% of the centile algorithm in early childhood.

A model like this needs further validation before we can apply it across the world, including gender differences and in different ethnicities. The great thing about growth charts is that they are a well known concept in paediatrics, and as Ian said, parents are already trained in the idea of monitoring growth in this way before they even see you. Once we have these validated for axial length, then that will probably be the tipping point for widespread implementation of axial length measurement in practice (start saving your pennies now for the instrument!). However if we have them validated for cycloplegic refraction and not axial length, perhaps axial length will continue to be an indicator for disease risk but still complicated for clinical tracking of myopia development and progression. Time will tell!

Caroline Klaver from Erasmus Medical Center in The Netherlands presented their protocol, which is based on growth curves they’ve previously published using large scale Dutch data. If a child presents with an axial length (AXL) on the 75th percentile or more for their age are prescribed 0.5% atropine and multifocal photochromatic spectacles. If a child presents with an AXL below the 75th percentile, they are offered low dose atropine, OrthoK or multifocal lenses according to the parents’ and patients’ suitability and wishes. If the child shows progression - doesn’t shift down to a lower percentile on the growth chart - then the treatment is altered to a higher dose of atropine or a combination treatment. All children are also advised to aim for two hours a day of outdoor exposure, and they’re reviewed every 6 months.

Caroline says they’ve received resistance from the Dutch Ophthalmology societies. The corneal society says you can’t do OrthoK, the risk of keratitis is too high! The paediatric ophthals say you can’t do higher dose atropine, the risk of phototoxicity is too high! Both recommend only low dose atropine as monotherapy. But that’s not stopped the good folk at Erasmus MC, and they’ve found a 70% reduction in myopia progression in 300 children over at least 3 years; and a 72% patient adherence to long term treatment.

Mark Bullimore and Noel Brennan have an important message to tell us about myopia control efficacy. This is borne from the difficulty in comparing percentage (%) efficacy rates across different studies - what does 56% or 38% actually mean for the patient in our chair, and how do interventions compare? The problem with percentages is that interventions can look less impressive (lower %) in younger children because as a whole they’re all progressing more quickly; and in older children the % can look impressively high because the control group is minimally progressing.

Alongside their colleague Xu Cheng, Mark and Noel analysed data from their own J&J experimental lens investigation (Cheng et al, OVS 2016) and to illustrate the problem with %, stratified the test and control groups by age. Over twelve months, the 8 year olds got a 34% reduction effect in their test lens; the 9 year olds got 25%; the 10 year olds got 42% and the 11 year olds 55%. BUT, all groups got the same statistical treatment effect of around 0.13mm absolute control effect.

In the next step, the team systemically reviewed 56 studies of myopia progression for absolute reduction in axial elongation (ARE). They found that most of the ARE occurred in the first year (mean 0.35mm), including the differences between more and less effective treatments - in the second year treatments were similar at a mean of 0.09mm of AXL control. Limited data was available to do a third year analysis, but the trend appeared similar.

The sum total of this is that:

  1. Xu, Noel and Mark propose the CARE factor - Cumulative Absolute Reduction of Axial Elongation. This means the total mm reduction achieved in a study, which also takes into account the duration of the study.
  2. The CARE of the available studies over a maximum of 2-3 years duration appears to be about 0.45mm or 1D. That doesn’t mean that we can’t achieve more over the longer term, just that this is the limit of what a global view of the data tells us at this point.
  3. Even if that sounds modest, Mark and Noel’s recent paper entitled ‘Myopia Control - why every diopter matters’ demonstrated that every D increases risks of myopic maculopathy (MM) by 67%, and that reducing the final amount of myopia by 1D should reduce the likelihood of MM by 40%. This relative risk and benefit holds right across the range, from 1 to 2D, 2 to 3D as well as from 5 to 6D

So the translation to clinical practice means that we should be cautious using predictors like the BHVI Myopia Calculator which potentially extrapolate 2-3 year study data over a whole childhood. This was also discussed above. Saying that we’ll be able to reduce a 6 year old’s future potential 8D of myopia to 5D might be overstating it. It’s useful to give parents an idea of what myopia progression means, but we should also explain the confidence intervals (potential variation for the individual).

The 1D message for reducing MM risk is an important one - Paul Gifford and I have included this in our new Myopia Management in Practice infographic which is available for free download from this website.. This infographic also includes a % efficacy message, but we’ve simplified this to 33% or 50%, so we can explain this to parents as that we can slow progression by ‘about a third’ or ‘about half’. These are easy concepts for parents to grasp. Read more on the research basis of the infographic here.

Sitting in this space between research and practice, constantly thinking about how to make it all work in the consulting room, I think that the ‘1D total efficacy’ message will sit best alongside robust predictors of future progression which include more data; and the ‘CARE’ factor will sit alongside axial length growth charts. Developments on both of these fronts have been presented at this meeting.

Kate profile thumbnail

About Kate

Dr Kate Gifford is a clinical optometrist, researcher, peer educator and professional leader from Brisbane, Australia, and a co-founder of Myopia Profile.

Leave a comment