Exploring the limits of myopia control efficacy

Paper title: Axial length targets for myopia control

Authors: Paul Chamberlain (1), Percy Lazon de la Jara (1), Baskar Arumugam (1), Mark A Bullimore (2)

1. CooperVision Inc., Pleasanton, USA
2. College of Optometry, University of Houston, Houston, USA

Date: May 2021

Reference:   Ophthalmic Physiol Opt. 2021;41:523-31 [Link to open access paper] 

Since pathology risk in myopic eyes is associated with excessive axial elongation,¹ it has been asserted that 'the goal of all clinical trials for myopia control should be the reduction of axial elongation'.² Whilst it is understood that both emmetropic and myopic eyes elongate throughout childhood, the amount of 'normal' axial elongation is still being determined, which influences how to determine ideal outcomes in myopia control. 

This novel analysis explored the absolute axial elongation of treated and untreated myopes in the MiSight 3-year clinical trial³ in comparison to previously published models of myopic and emmetropic eye growth. The results demonstrate that 3 year axial elongation in MiSight treated myopes approached that of 'virtual' age-matched cohorts of emmetropes. The 'untreated' (control) MiSight myopes showed similar axial elongation to the historical models, validating the comparison.

Since emmetropic eyes elongate in childhood, these results indicate a potential limit to the short-term percentage efficacy of myopia control treatment, as a 'physiological' component of eye growth can be expected to be retained. 

Annual axial elongation is higher in myopes than in emmetropes, and annual axial elongation decreases with age. 

• Previous large scale, longitudinal studies of childhood refractive development (OLSM4 and SCORM5 studies) have shown that axial elongation occurs in eyes which maintain emmetropia. This data was used as a model to compare with the MiSight 3-year clinical trial results.³
• Since axial elongation in childhood is a normal state in emmetropia, it cannot be expected to be stopped in myopia control. In emmetropes, flattening and thinning of the crystalline lens counteracts axial elongation to maintain refractive error, while in myopes the lens fails to compensate for the increasing axial length.  

The mean 3-year elongation in MiSight treated myopes is less than half of the progression for untreated control subjects (those wearing Proclear single vision contact lenses). This is presumed to indicate a comparative treatment efficacy of around 50% over three years.³

• The progression of MiSight-wearing myopes was not substantially different from the predicted axial elongation rates for the virtual cohorts of emmetropes (0.24mm over 3 years, for both OLSM and SCORM models). Could this be considered a greater-than-50% effect if the axial elongation was slowed to effectively 'normal' rates? 
• Clinically, this complicates determining success in myopia control treatments. Since even emmetropic eyes will grow, Eye Care Practitioners should not anticipate a total cessation of eye growth in myopia control. 

The untreated control myopes in the MiSight clinical trial show similar cumulative axial elongation over 3 years (0.62mm) to those predicted for the virtual cohorts of myopic children based on both the OLSM (0.70mm) and SCORM (0.63mm) models.

• This supports the reliability and expands the applicability of the results obtained in the MiSight 3-year clinical trial.³

• Firstly, the older OLSM and SCORM comparative studies utilized contact ultrasound measurement of axial length. This was typical at the time, but reduces the accuracy of the axial length measurements by around ten times² compared to axial length measurement by optical biometry in the recent MiSight 3-year clinical trial.³
• Differences in the cohorts of the previously published growth models and the MiSight 3-year participants. This may contribute to discrepancies between the virtual cohorts derived from the models. 
  • For example, the OLSM model included children aged 6-14 years of age, SCORM included children aged 6-12 years and the MiSight study 8-15 years. There were no comparative emmetropes in the MiSight study, and different proportions of myopes in the SCORM and OLSM studies. Ethnicity was similar in OLSM (USA based, predominantly Caucasian) and MiSight (Canada, Portugal, Singapore, United Kingdom) cohorts which was dissimilar in SCORM (Singaporean children). Interestingly, though, the SCORM myopes appear to have greater annual axial elongation at age 8 and 9 but slightly lesser from age 10, and both groups of emmetropes appear similar. 
  • Vitreous chamber depth (VCD) was used as a substitute for axial length in the SCORM model as the axial length equations asymptotes at age 13 years. The authors indicate that this could have slightly under-estimated SCORM axial changes. 
  • It is noteworthy that the OLSM (1989-2001) and SCORM (1999-2002) data was gathered 2-3 decades ago while the MiSight 3-year study took place from 2012-2017. Despite acknowledged worldwide and country-specific increases in myopia incidence rates in the last 2-3 decades, it is fascinating that this comparison indicates that after onset, myopic eye growth appears similar in current versus historical myopes. The paper goes on to make comparisons with a handful of other recent clinical trials which indicate similar axial elongation in untreated progressive myopia.
• It isn't possible to say whether the OLSM, SCORM and MiSight groups are statistically similar or otherwise; rather that they 'look similar' or 'approach each other'. The data is compelling and a novel view on myopia control efficacy, but the authors themselves caution against adjusting myopia control efficacy to remove an 'emmetropic correction factor'. Their reasons are that absolute, not relative, axial growth is of key importance to ocular health and that there's currently no evidence for a 'pathological' growth mechanism in myopia as distinct from the 'physiological' mechanism in emmetropia. While we are still learning about how to best express and clinically judge myopia control efficacy, making consistent comparisons is essential.
• The analysis represents an average change across a population of children, and as is the case in any myopia control study, the individuals within the population displayed great variation. The standard deviations for axial elongation for the 3-year MiSight trial were considerable, representing a range of individual outcomes (axial elongation for treated group 0.30 ± 0.27mm; control group 0.62 ± 0.30mm). Further research is required to investigate contributing factors to the wide variation in observed treatment effect.

The goals of this analysis were to firstly compare eye growth measured in treated and untreated myopes in the MiSight 3 year clinical trial to untreated myopes and emmetropes in previously published population studies of childhood eye growth. The second was to utilize large data sets on normal physiological eye growth from previous studies to characterize the myopia control effect found with MiSight and to appropriately set myopia control expectations.

The study utilised axial elongation data obtained during a 3-year RCT of MiSight 1 Day.³ Myopic children (n=144)  aged 8-12 years were enrolled in the study and were randomised into either a MiSight contact lens (treated) or Proclear 1 day single vision contact lens (untreated). Changes in refractive error and axial length were highly correlated. Over three years, unadjusted mean myopia progression was 0.73D lower in the treated group than the untreated group (-0.51 ± 0.64D vs -1.24 ± 0.61D, p < 0.001). Mean axial elongation was 0.32 mm less in the treated group than the untreated group (0.30 ± 0.27mm vs 0.62 ± 0.30 mm, p < 0.001). 

This longitudinal axial length data was then compared to data for myopic and emmetropic children in two large cohort studies. These were firstly the Orinda Longitudinal Study of Myopia (OLSM)⁴ in the US, a twelve-year study which included 737 children between 6-14 years of age. Of these children, 247 became myopic, where 76% were emmetropic or hyperopic at baseline, and 194 were persistent emmetropes.

The second comparator was the three-year Singapore Cohort Study of the Risk Factors for Myopia (SCORM)⁵ study, which included 1775 children (mainly Chinese) between 6-12 years of age. Of these children, 616 were progressing myopes and 369 were emmetropes. Vtireous chamber depth (VCD) data was used for this cohort due to age limitations past age 13 in the axial length data. 

Both studies published equations for eye growth in emmetropes and progressing myopes^6,7 that were subsequently compared with the MiSight clinical trial data. Using these previous published equations, annual axial (or VCD) elongation for ages 8-14 years was calculated. The OLSM and SCORM myopes and emmetropes were found to be broadly similar.  

Then, a 'virtual cohort' of myopes and emmetropes from each of the OLSM and SCORM models were created. Each of these four virtual cohorts had the same age distribution as the MiSight clinical trial subjects at baseline (mean 10.1 years). The predicted elongation was calculated at years 1, 2 and 3 for myopes and emmetropes using both the OLSM and SCORM models.

The results shown in Figure 1 from the paper indicate that 'untreated' myopes in the MiSight study (wearing the control single vision contact lens) showed similar axial elongation over each of three years to the single-vision wearing myopes in the OLSM and SCORM matched cohorts. This validated that the MiSight untreated myopes were similar to historical myopes. 

The most interesting finding is that the MiSight-wearing myopes ('treated') showed similar eye growth to the historical emmetropes, indicating a slowing of eye growth to 'physiological' levels expected in emmetropic children. This led to the formation of two hypotheses by the authors regarding eye growth - that myopic axial elongation includes a component of physiological growth, and that optical myopia control strategies (or at least, in the case of MiSight) may minimize the former but not prevent the latter.

When using short-term percentage comparisons, the OLSM and SCORM emmetropes show around 30-40% rate of axial elongation to (untreated) myopes. Hence, 'controlling myopia' to the limit of only observing the emmetropic rate of eye growth would predict a ceiling of 60-70% comparative efficacy. Perhaps this indicates the physiological limit of myopia control efficacy - can the myopic eye be expected to grow more slowly than an emmetropic eye? Is this a reasonable goal if absolute eye length is the key ocular health risk factor? Future comparison of treatment groups to matched emmetropes, and even concurrent emmetropic groups in myopia control studies, could help elucidate this potential efficacy maximum. 

This paper demonstrates that 'untreated' (control) myopes in the MiSight 3-year clinical trial demonstrated axial elongation similar to that predicted in previously published models of eye growth (SCORM AND OLSM). Myopic children 'treated' by wearing MiSight demonstrated axial elongation similar to that of emmetropes predicted by models of eye growth. Eye care practitioners should remain cognisant of current research in both untreated progressive myopic and emmetropic eye growth, in order to set realistic expectations for myopia control efficacy.

Title: Axial length targets for myopia control

Authors: Paul Chamberlain, Percy Lazon de la Jara, Baskar Arumugam, Mark A Bullimore

Purpose: Both emmetropic and myopic eyes elongate throughout childhood. The goals of this study were to compare axial elongation among untreated progressing myopes, progressing myopes treated with a myopia control contact lens and emmetropes, in order to place axial elongation in the context of normal eye growth in emmetropic children, and to consider whether normal physiological eye growth places limits on what might be achieved with myopia control.

Methods: Axial elongation data were taken from the 3-year randomised clinical trial of a myopia control dual-focus (MiSight® 1 day) contact lens. These were compared with data for myopic and emmetropic children in two large cohort studies: the Orinda Longitudinal Study of Myopia (OLSM) and the Singapore Cohort Study of the Risk Factors for Myopia (SCORM). Each study's published equations were used to calculate annual axial elongation. Four virtual cohorts-myopic and emmetropic for each model-were created, each with the same age distribution as the MiSight clinical trial subjects and the predicted cumulative elongation calculated at years 1, 2 and 3 for myopes and emmetropes using both the OLSM and SCORM models.

Results: The untreated control myopes in the MiSight clinical trial showed mean axial elongation over 3 years (0.62 mm) similar to the virtual cohorts based on the OLSM (0.70 mm) and SCORM (0.65 mm) models. The predicted 3-year axial elongation for the virtual cohorts of emmetropes was 0.24 mm for both the OLSM and SCORM models-similar to the mean 3-year elongation in MiSight-treated myopes (0.30 mm).

Conclusions: The 3-year elongation in MiSight-treated myopes approached that of virtual cohorts of emmetropes with the same age distribution. It is hypothesised that myopic axial elongation is superimposed on an underlying physiological axial elongation observed in emmetropic eyes, which reflects increases in body stature. We speculate that optically based myopia control treatments may minimise the myopic axial elongation but retain the underlying physiological elongation observed in emmetropic eyes.