Effect of peripheral defocus on axial growth and modulation of refractive error in children with anisohyperopia

To establish whether axial growth and refractive error can be modulated in anisohyperopic children by imposing relative peripheral hyperopic defocus (RPHD) using multifocal soft contact lenses.

Until recently, 16 attempts to modulate refractive error in children with hyperopia have been absent from the literature and remain unaddressed in those with anisohyperopia.
Stochastic factors appear to play a role in the aetiology of refractive errors as inferred by the existence of anisometropia. 6 In anisometropes, despite sharing the same genome, and for the most part, similar environmental exposure, each eye emerges with a different refractive error. Anisometropia can be present early in life with subsequent ocular development complicated by amblyopia. 6,7 However, anisometropia often develops later; it seems that the regulated pattern of growth at around the age of 6 years exhibits increasing variability between individuals and between the two eyes of an individual. 6 The ability to modulate axial growth in a range of species by imposing single vision, full-field relative hyperopic and myopic defocus has been widely reported. [17][18][19][20][21] Furthermore, the literature demonstrates that short-term changes in AL and choroidal thickness occur in response to hyperopic defocus in humans. [22][23][24][25][26] The progression of myopia and axial growth can be slowed in children and adolescents using soft multifocal or dual-focus contact lenses. [27][28][29][30] These contact lenses are designed to impose myopic defocus through the outer optic zone, while simultaneously correcting distance refractive error through the central optic zone. More recently, work from the present authors has shown the ability to accelerate axial growth in children with isohyperopia using centre-near multifocal contact lenses to impose relative peripheral hyperopic defocus (RPHD). 16 The ability to modulate refractive error and axial growth in children with anisohyperopia by imposing RPHD has not yet been tested. Given that axial growth can be accelerated in children with isohyperopia by imposing hyperopic defocus using multifocal contact lenses, 16 applying a similar principle to anisohyperopes is a natural extension to this work.

METHODS
Prior to commencing the research, ethical approval was obtained from both the National Health Service Health Research Authority and Aston University Research Ethics Committees with the study designed to follow the tenets of the Declaration of Helsinki. Each participant, and their parent or guardian where appropriate, was given detailed information regarding the nature of the study, both verbally and in written form; this allowed informed consent and assent to take place prior to participation. Participants were required to complete a short questionnaire 31 to ensure they met the inclusion criteria. The programme of research was registered as a clinical trial: ClinicalTrials.gov NCT02686879. Suitable candidates for the study were recruited by displaying notices at the research sites. Potential participants were also sourced through a database search at the research venues to identify individuals who met the age and refractive error inclusion criteria.
Participants aged between 8 and <16 years were recruited. Axial growth and refractive error were observed without intervention for the first 6 months of the trial with participants wearing their habitual spectacle correction during this period. Between the 6-and 30-month time points of the 3-year trial, participants wore a centre-near multifocal soft contact lens in their more hyperopic eye for a minimum of 10 h per day for 6 days per week. A single vision contact lens was worn in the fellow eye if required, that is if the level of hyperopia in this eye required refractive correction. Monthly disposable, comfilcon A multifocal contact lenses (Biofinity, coope rvisi on.com) with a centrenear design and a +2.00 D add were worn throughout the intervention period in the more hyperopic (intervention) eye and an equivalent single vision contact lens (monthly disposable, comfilcon A) was worn in the fellow (control) eye if required. The power of the central portion of the intervention lens was selected to correct distance refractive error while simultaneously exposing the retina to RPHD from the outer distance zone (see Figure 1). A +2.00 D add was selected in line with previous refractive error modulation studies 27,28,32 to strike a balance between ensuring adequate visual performance 33 while imposing peripheral defocus at a level sufficient to test the hypothesis. 30 The power profile of the Biofinity centre-near lens design has a measured add close to its nominal value at +1.83 D for +2.00 D. 34 The intervention was withdrawn for the remaining 6 months of the trial and participants reverted to optimal single vision spectacle lens correction as determined at the penultimate visit using standard objective and subjective refraction. Changes to axial growth and refractive error were measured at the final visit.
Inclusion and exclusion criteria are summarised in Table 1 and are in line with earlier work in children with isohyperopia. 16 Full details of primary and secondary outcomes' measures, including time points, are provided in Tables 2-5 and Figures 2 and 3. Key parameters

Key points
• Unlike similar work with isohyperopes, the rate of eye growth did not increase using the multifocal soft contact lens specified in the present study for children with anisohyperopia. • The reduction in refractive error did not accelerate using the multifocal soft contact lens specified in the present study for children with anisohyperopia. • Earlier intervention and imposing a greater degree of peripheral defocus is an opportunity for future work to build upon the findings in the present study.
F I G U R E 1 Schematic to demonstrate the concept of relative peripheral hyperopic defocus imposed in the more hyperopic eye with a centrenear multifocal contact lens (CL) while the full refractive error is corrected centrally. 16 T A B L E 1 Summary of inclusion and exclusion criteria. Intervention eye (n = 11) Control eye (n = 11)

Inclusion criteria Exclusion criteria
Note: The intervention period is shaded orange.
Abbreviation: SEM, standard error of the mean.
included measures of unaided distance vision and distance visual acuity (DVA) at 6 m along with near visual acuity at 0.25 m undertaken with high contrast logMAR charts and determined using a by-letter scoring method (0.02 logMAR units per letter). Biometric assessment included measures of AL, anterior chamber depth (ACD) and corneal curvature (CC) and was taken using the IOLMaster 500 (zeiss.com). 35 For AL, 10 measurements were taken per eye and the composite value recorded. Subjective refraction was recorded prior to instillation of cyclopentolate hydrochloride 1% using standard optometric techniques. Measures of accommodative lag were also obtained (prior to cycloplegia) using the Grand Seiko WAM-5500 autorefractor (grand seiko. com). 36 Participants were asked to view a high-contrast Maltese cross, 25 mm in size, binocularly at a distance of 0.33 m while wearing their contact lens correction, with measures of accommodative lag taken from the dominant eye only, 37 which was determined with the commonly used hole-in-the-card test to achieve a binary outcome. 38 Amplitude of accommodation was assessed using a Royal Air Force rule with the mean of three push-up and three pull-down measures reported. 39,40 Objective central refraction was measured 30 min after instillation of the cycloplegic agent using the Grand Seiko WAM-5500 autorefractor while viewing a diffuse target at a 6 m equivalent distance. Post-cycloplegic peripheral refraction measures were undertaken using the same instrument at 30° temporally, 30° nasally, 20° superiorly and 20° inferiorly. Here, participants were asked to fixate on high-contrast Maltese crosses, 25 mm in size, in photopic conditions (440 lux) which were placed on a wall at 1.64 m to achieve the desired eccentricity for each of the four peripheral measures. Non-cycloplegic measures of peripheral refraction were also taken with contact lenses in situ. Central contrast sensitivity was recorded monocularly with spectacle and contact lens correction using a computerised version of the Pelli-Robson chart (thoms on-softw are-solut ions.com) at a distance of 1 m. Stereoacuity was measured using the TNO Randot Stereotest (Edition 15, lamer is-group.nl) at a distance of 0.4 m with the spectacle and contact lens correction.

Statistical analyses
All data were analysed using commercially available software, SPSS, v. 25, (ibm.com). Sample size calculation indicated that a total of 11 participants would be required to achieve 80% power for an effect size of 0.25 at a significance level of 5% using a mixed factor repeated measures ANOVA design (G*Power 3.1, psych ologie.hhu.de/arbei tsgru ppen/allge meine -psych ologi e-und-arbei tspsy cholo gie/gpower). Data were examined with the Bonferroni correction applied throughout. [41][42][43]

R ESULTS
In total, 11 participants were recruited, comprising eight females and three males with an age range at baseline of 8.25-13.42 years, mean 10.56 years (SD 1.43 years); these data were normally distributed (Kolmogorov-Smirnov, Z = 0.18, p = 0.20). All participants completed the trial and reported compliance with the prescribed wearing schedule. There were no adverse events related to contact lens wear. All participants required contact lens correction in both eyes. The primary outcome measures were changed to AL and post-cycloplegic central refractive error which are detailed in Tables 2 and 3 and Figures 2  and 3. Secondary outcome measures are summarised in Tables 4 and 5. Overall, AL changed over time (F (6,60) = 14.81, p < 0.001), although an interaction between factors demonstrated that this occurred in the control eye only (F (6,60) = 2.61, p = 0.03). For the intervention eye, AL did not change from baseline to the 6-month time point (p > 0.99), throughout the 2 years of intervention (p = 0.06) or after the intervention was withdrawn for the final 6 months of the trial (p > 0.99). For the control eye, AL did not change significantly from baseline to the 6-month time point (p = 0.05). Axial growth accelerated throughout the 2 years of intervention (p = 0.003) but did not change once the intervention was withdrawn for the final 6 months of the trial (p > 0.99). Observed power was 0.82. AL data from baseline to the end point of the  trial are detailed in Table 2, with changes over time illustrated in Figure 2.
Post-cycloplegic mean spherical equivalent central refractive error decreased over time (F (4,40) = 4.60, p = 0.004) with an interaction between factors demonstrating that this was similar in both the intervention eye and the control eye (F (4,40) = 0.32, p = 0.86) and the observed power was 0.11. Refractive error data from baseline to the end point of the trial are given in Table 3, with changes over time provided in Figure 3.

DISCUSSION
This paired-eye clinical trial has explored for the first time whether the interocular differences in AL and refractive error at baseline can be reduced in children with anisohyperopia by imposing RPHD unilaterally in the more hyperopic eye, with the fellow eye serving as a control. Measurements of peripheral refraction with the multifocal contact lens in situ showed that the intended level of RPHD was achieved in the intervention eye. Nevertheless, despite the imposition of RPHD, AL did not increase significantly over time. Intriguingly, the AL in the control eye did increase over the trial period, and in fact, the interocular difference in AL was greater at the point of exit from the trial than at baseline. The findings here are in contrast to similar work with isohyperopes 16 where axial growth accelerated with the imposition of RPHD.
While mean post-cycloplegic refractive error decreased over time, this was by a similar amount in both the intervention and control eyes. As with AL, there was a failure to close the gap between the interocular difference in refractive error measured at baseline, although this outcome measure did not achieve statistical power. It is also worth observing the variability associated with measurement of refraction, 44 which may be a factor in this outcome. The lack of response to the imposition of RPHD may be due to a failure of eyes with high levels of hyperopia to emmetropise early in life; a theory postulated in earlier work where F I G U R E 2 Change in axial length (mean ± SEM). these hyperopes are 'left behind' rather than being regulated towards emmetropia. 6 Stereoacuity was similar to both spectacles and contact lenses, which is in keeping with findings from earlier work showing that stereoacuity appears to be preserved in multifocal contact lens wear compared to single vision correction, albeit in a presbyopic cohort. 45 In terms of contrast sensitivity, measures were similar to spectacles compared with both single vision and multifocal contact lenses and did not change over time; this offers reassurance that visual performance appears adequate for young wearers with this form of correction.
With respect to anterior eye parameters, as with previous refractive error modulation work in myopes, 29,46 CC did not change over time in either eye. Similarly, ACD did not change over time suggesting that the longitudinal increase in AL observed in the control eye was likely to be due to vitreous chamber depth changes.
Accommodative lag with spectacle correction in the intervention eye did not change over time although it was significantly less than with contact lenses, which may reflect the fact that hyperopes would be expected to converge less through the former mode of correction. The finding here is interesting given that impaired lag has been implicated as a driver for axial growth progression in myopes; 47,48 nevertheless, this does not appear to be the case here.
As with previous work, [49][50][51][52][53][54] peripheral refraction was relatively myopic in all four quadrants in both the intervention and control eyes and did not change over time. Importantly, the trial demonstrated that centre-near multifocal contact lenses offer a viable method to induce defocus in anisohyperopes compared to the control eye fitted with a single vision contact lens. However, the role of peripheral defocus in the manipulation of axial growth remains uncertain, and the primary outcomes observed here for anisohyperopes fail to add clarity.
All participants completed the trial, which supports the findings from earlier work that children can successfully and safely transition into contact lens wear 55 and also adapt well to wearing a multifocal design unilaterally. While participants appeared to comply with the wearing regime outlined in Table 1, this was a self-reported measure.
While the imposition of RPHD appears to influence axial growth in isohyperopes, 16 the same does not hold true for anisohyperopes using the paradigm outlined in the present work. It seems that, unlike isohyperopes, the unique growth patterns typically experienced by anisohyperopes 7 are resistant to the influence of RPHD, at least in this age demographic and at the magnitude of defocus tested here. In the present trial, a single power, 'off-the-shelf' presbyopic design was used. Intervention with a tailored design using a smaller central zone and higher add power to impose greater levels of peripheral defocus may have greater impact on the modulation of axial growth and refractive error in these children.
For anisohyperopes, instinctively, the eye closer to emmetropia (the control) would be regarded as the 'normal' eye. However, given that the mean growth rate for the more hyperopic eye in the present work was closer to the expected norm, 56 should the control eye be considered as the 'abnormal' one of the pair? With this in mind, perhaps the primary endeavour should be to slow down growth in the least hyperopic eye, rather than attempting to accelerate growth in the more hyperopic eye. Taking this further, imposing competing defocus models could potentially yield the greatest result for anisohyperopes; that is to say, using myopic defocus to slow down growth in the least hyperopic eye, with the opposite approach taken in the fellow eye.
It remains to be seen whether earlier intervention and taking a more aggressive approach to defocus would yield more promising results; this provides an opportunity for future work to help avoid the near lifelong visual impairment that these individuals otherwise face.

AC K N O W L E D G E M E N T S
This research was funded by a postgraduate scholarship from the College of Optometrists.

CO N F L I C T O F I N T E R E S T S TAT E M E N T
The authors confirm that they do not have any commercial interest relevant to the subject of the paper. The contact lenses in the study were supplied free of charge by CooperVision. CooperVision did not sponsor the research and does not support or have an opinion regarding any of the content in the paper.