Peripheral Defocus, Pupil Size, and Axial Eye Growth in Children Wearing Soft Multifocal Contact Lenses in the BLINK Study

Abstract

Purpose: The purpose of this study was to evaluate the relationship between peripheral defocus and pupil size on axial growth in children randomly assigned to wear either single vision contact lenses, +1.50 diopter (D), or +2.50 D addition multifocal contact lenses (MFCLs).

Methods: Children 7 to 11 years old with myopia (−0.75 to −5.00 D; spherical component) and ≤1.00 D astigmatism were enrolled. Autorefraction (horizontal meridian; right eye) was measured annually wearing contact lenses centrally and ±20 degrees, ±30 degrees, and ±40 degrees from the line of sight at near and distance. Photopic and mesopic pupil size were measured. The effects of peripheral defocus, treatment group, and pupil size on the 3-year change in axial length were modeled using multiple variables that evaluated defocus across the retina.

Results: Although several peripheral defocus variables were associated with slower axial growth with MFCLs, they were either no longer significant or not meaningfully associated with eye growth after the treatment group was included in the model. The treatment group assignment better explained the slower eye growth with +2.50 MFCLs than peripheral defocus. Photopic and mesopic pupil size did not modify eye growth with the +2.50 MFCL (all P ≥ 0.37).

Conclusions: The optical signal causing slower axial elongation with +2.50 MFCLs is better explained by the lens type worn than by peripheral defocus. The signal might be something other than peripheral defocus, or there is not a linear dose-response relationship within treatment groups. We found no evidence to support pupil size as a criterion when deciding which myopic children to treat with MFCLs.

The global prevalence of myopia is increasing.¹ The risk of comorbidities (such as myopic macular degeneration, retinal detachment, glaucoma, etc.) due to excessive axial elongation increases as myopia progresses, as does the economic impact and burden of healthcare associated with increases in myopia prevalence and severity.^1–5 There is evidence across animal species that eye development is visually guided, with the eyes of young animals increasing or decreasing the rate of axial elongation to match the sign and amount of lens-induced defocus.^6–8 Visually guided mechanisms regulating eye growth operate in a regionally specific manner in animal models with only the portion of the eye experiencing a visual signal changing growth in response.^9,10 However, it is not clear that this is the case in children based on our longitudinal myopia control peripheral eye growth data that found symmetric eye elongation across the periphery rather than enhanced peripheral inhibition in those wearing multifocal contact lenses in the Bifocal Lenses in Nearsighted Kids (BLINK) Study.¹¹ Animal experiments involving lenses with a clear central aperture that only produce peripheral hyperopia and normal central vision still result in accelerated axial eye growth, demonstrating that the peripheral retina can guide refractive error development.¹² Bi-directional changes in choroidal thickness have also been reported in response to the sign of defocus, providing further evidence that the eye detects the type of retinal blur.^13–15 

Based on these results in animal models, optical treatment strategies that impose myopic retinal defocus have been investigated in children as a method to slow the progression of juvenile-onset myopia.¹⁶ Although clinical trials evaluating progressive addition spectacle lenses found statistically significant but clinically small reductions in myopia progression,^17,18 this type of lens primarily causes myopic defocus in the superior retina as opposed to throughout the periphery.¹⁹ Whereas spherical soft contact lenses can cause changes in peripheral defocus,^20,21 more significant changes are caused by orthokeratology²² and rotationally symmetric center-distance soft multifocal contact lenses,^23–25 which create myopic defocus throughout the retinal periphery while maintaining good visual acuity. 

Randomized clinical trials ranging from 2 to 3 years in duration evaluating orthokeratology^26,27 and soft multifocal contact lenses^28,29 have had success in slowing myopia progression and eye growth compared to single vision corrections. Spectacle designs that incorporate plus power throughout the lens outside of a central clear zone have also had success in slowing myopia progression.^30,31 These optical treatments are hypothesized to slow eye growth by imposing myopic defocus on the retina; however, testing of this hypothesis is needed. Larger pupil size is also hypothesized to enhance the myopia control effect by exposing more of the retina to the plus power in these optical treatments, thereby enhancing the myopia control effect.^32–34 

The BLINK Study found children wearing +2.50 add center-distance multifocal contact lenses had less axial growth over 3 years compared to children wearing single vision contact lenses, but children wearing +1.50 add multifocal contact lenses did not have slower progression.²⁹ The purpose of this analysis was to determine whether peripheral defocus created by multifocal contact lenses was associated with the 3-year reduction in axial elongation. We also investigated whether pupil size modified the slowing of axial elongation with multifocal contact lenses. 

Methods

Two hundred ninety-four children with myopia aged 7 to 11 years (inclusive) were enrolled in a double-masked, 3-year, randomized clinical trial at 2 sites (University of Houston and Ohio State University Colleges of Optometry) between September 2014 and June 2016. The study protocol was approved by the institutional review board at each institution. This research adhered to the tenets of the Declaration of Helsinki, was monitored by an independent Data and Safety Monitoring Committee, and was supported by the National Eye Institute. Prior to screening for enrollment, children provided assent and parental permission was obtained from the child's parent or guardian. Eligible participants had +0.10 logMAR or better best-corrected high-contrast distance visual acuity in each eye, −0.75 to −5.00 D of myopic refractive error in each meridian (inclusive), 1.00 D or less astigmatism in each eye, and 2.00 D or less anisometropia as measured by cycloplegic autorefraction. Full details of the participant baseline characteristics and study methods have been previously reported.^29,35 Methodological details pertinent to this analysis are described below. 

Children were randomly assigned to wear Biofinity single-vision soft contact lenses, Biofinity Multifocal D contact lenses with a +1.50 D add, or Biofinity Multifocal D with a +2.50 D add power (CooperVision; Pleasanton, CA, USA). In vitro power profiles of the multifocal lenses used in this study have been previously reported.^36,37 Briefly, the Biofinity Multifocal D is a center-distance multifocal design with a central distance zone with no add power from the lens center out to a diameter of about 3.2 mm, an intermediate region in which there is a progressive increase in plus power beginning from a lens diameter of approximately 3.2 mm out to 4.2 mm, and then a near zone of plus power from a diameter of 4.2 mm to 8 mm (the edge of the optic zone). The measured power profile of the Biofinity Multifocal D with +2.50 add was reported to have maximum plus power that varies depending on the distance power. Lower minus lenses had maximum plus power that was maintained from the mid-periphery out to the edge of the optic zones, whereas higher minus (more negative power) lenses had the most plus power in the mid periphery with some decrease in plus toward the edge of the optic zone with lenses that corrected the greatest amount of myopia.³⁶ 

Axial length was measured at baseline and annually after cycloplegia using the Lenstar LS 900 (Haag-Streit USA, Mason, OH, USA). Five measurements at each visit were averaged. Cycloplegia was achieved by 2 drops of 1% tropicamide, separated by 5 minutes. Outcome measures made under cycloplegia commenced 30 minutes after the first drop of tropicamide. 

Central and Peripheral Defocus

Autorefraction measurements of the right eye out to ±40 degrees from the line of sight were made using a modified Grand Seiko WAM-5500 autorefractor (Visionix USA, Bensenville, IL, USA).³⁵ For distance measurements, a laser diode emitting red light was placed in holding wells mounted on top of the instrument that were positioned to allow for measurements in 10 degree increments from the line of site. The same laser holding wells also allowed for a near point rod to be placed at the same locations for the purpose of near measurements. These measurements were made annually while wearing the study contact lens looking at near (before cycloplegia) and at distance (after cycloplegia) along the line of sight and ±20, ±30, and ±40 degrees from the line of sight in the horizontal meridian while the left eye was patched. Participants turned their head, not their eye, for all peripheral measurements. The examiner ensured an appropriate head movement was made with the eye remaining in primary gaze and that there was no head tilt before each location was measured. The instrument reticule was centered within the entrance pupil³⁸ while five valid measurements (within 1.00 D of the median for sphere and cylinder) were performed, and each measurement was converted to vector space (M, J₀, and J₄₅) before averaging.³⁹ 

The near target for each location measured was a grid of 20/125 letters at 33 cm (3 diopter demand). The near defocus profile experienced by the retina was calculated by adding the 3-diopter target demand to the mean spherical equivalent defocus (M) measured at each location. Participants looked at a laser spot projected on the wall for each distance location measured. The study contact lens worn for near measurements was the habitual power the participant came in wearing to the annual visit, representing what the eye most recently experienced; at baseline, it was the initial contact lens power dispensed. The contact lens power worn for distance measures was the power being dispensed at the annual visit representing future exposure with optimal correction. 

Autorefraction measurements while wearing contact lenses from the baseline, 1-year, and 2-year visits were used to calculate multiple metrics to evaluate whether there was an association between retinal defocus and the 3-year change in axial length. For each participant, spherical equivalent defocus was averaged across the baseline, year 1, and year 2 visits to create defocus metrics that were evaluated in models. 

Aggregate Peripheral Defocus at Distance and at Near

The aggregate metrics of defocus included all peripheral locations measured in a single variable. These metrics assume that all retinal locations contribute collectively to the overall growth of the eye and treatment effect based on observations from animal models that local visual signals influence local ocular growth.^9,10 Peripheral defocus at distance was defined as the mean defocus averaged across all visits and peripheral locations while looking at distance, creating a single metric for each child. Peripheral defocus at near was calculated in the same way as the mean defocus at near averaged across all visits and peripheral retinal locations. 

Most Myopic Meridian at Distance

In rhesus monkeys exposed to astigmatism, emmetropizing eyes generally grow toward the least hyperopic (most myopic) of the two image planes rather than to the circle of least confusion.⁴⁰ Astigmatism increases in the periphery of the eye, so a metric was created to quantify the most myopic meridian (focal plane) across the horizontal retina while wearing study contact lenses. For each participant, the most myopic meridian was calculated at each retinal location by adding the spherical and cylindrical value from each autorefraction measurement at that location (in minus-cylinder form) and then averaging those values. The meridian with the most myopia across the locations measured was determined. These most myopic meridian values were averaged across visits for each participant, and this metric was evaluated as a predictor. 

Central Defocus at Near

The central retinal defocus value when looking at near was evaluated wearing the habitual contact lens worn to the visit, which measured the defocus experienced at the fovea resulting from the optics of the contact lens combined with the child's level of accommodation. Central defocus at near was averaged across visits and evaluated as a predictor. 

Defocus at Distance by Retinal Location

Finally, to address the possibility that some locations on the retina are more influential than others, each child's mean defocus across visits at each retinal location while looking at distance was calculated. These location-specific defocus metrics were each evaluated as a predictor in models. 

Pupil Size

Pupil size (right eye) was measured annually under mesopic (approximately 2 lux) and photopic (approximately 500 lux) conditions to the nearest 0.1 mm using a NeurOptics VIP-200 Pupilometer (NeurOptics, Inc., Irvine CA, USA) while the other eye fixated at distance. The mean pupil size averaged over all visits was calculated for both mesopic and photopic conditions and included in the models. 

Models

Models of the 3-year change in axial length were fitted that included the defocus metrics described above in addition to sex at birth, study site, age, baseline axial length, and mesopic pupil size. Pupil size in mesopic lighting was selected to determine whether any effect of defocus on axial elongation with multifocal contact lens was enhanced by having a larger pupil. Interactions between each defocus metric and mesopic pupil size were explored. Models were repeated including treatment group to determine whether the 3-year change in axial length was associated with defocus within each treatment assignment. Interactions between defocus metrics and treatment group were evaluated to determine the consistency of any association across treatment groups. Interactions between pupil size (both mesopic and photopic) and treatment group were also evaluated to determine whether pupil size influenced the slowing of eye growth with the +2.50 multifocal previously reported in the BLINK Study.²⁹ 

Results

A total of 287 children out of 294 enrolled completed a 3-year visit allowing for the calculation of the 3-year change in axial length. Baseline demographic and ocular characteristics by treatment group have been previously published with the study primary outcome results.²⁹ At enrollment, the mean (± SD) age, spherical equivalent refractive error (right eye), and axial length were 10.3 ± 1.2 years, −2.39 ± 1.01 D, and 24.49 ± 0.81 mm, respectively. The baseline mean (± SD) right eye spherical equivalent refractive error in the +2.50 add group, +1.50 add group, and single vision group were −2.28 ± 0.90 D, −2.43 ± 1.11 D, and −2.46 ± 0.97 D, respectively. The 3-year change in axial length and the average mesopic and photopic pupil size are in Table 1. As reported previously, axial growth was −0.23 mm (95% confidence interval [CI] = −0.33 to −0.14 mm) less in the +2.50 add group than in the single vision group (P < 0.001), was −0.16 mm less (95% CI = −0.26 to −0.07 mm) in the +2.50 add group than the +1.50 add group (P = 0.002), and was not different between the +1.50 add and single vision groups (−0.07 mm; 95% CI = −0.16 to 0.03 mm; P = 0.15).²⁹ There were no significant differences in pupil size among treatment groups (mesopic, P = 0.66 and photopic, P = 0.51). The mean (± SD) mesopic and photopic pupil size across treatment groups over 3 years were 6.4 ± 0.7 mm (range = 4.5 to 8.4 mm) and 5.0 ± 0.6 mm (range = 3.3 to 6.9 mm), respectively. 

Table 1.

 

Mean (± SD) 3-Year Change in Axial Length, Mesopic Pupil Size, and Photopic Pupil Size by Treatment Group

 

The mean defocus profile while wearing contact lenses averaged across the baseline, 1-year, and 2-year visits are shown while looking at distance (Fig. 1A) and at near (Fig. 1B). At distance, multifocal contact lenses resulted in more myopic peripheral defocus relative to single vision contact lenses that increased with the add power (i.e. the +2.50 add showed less peripheral hyperopia or more peripheral myopia than the +1.50 add at all locations; all P < 0.008). The smaller myopic changes of the +1.50 multifocal only resulted in myopic retinal defocus on average at two of the retinal locations measured (−0.08 D at 20 degrees nasal and −0.20 D at 20 degrees temporal). The larger myopic changes caused by the +2.50 multifocal resulted in myopic defocus on average at 4 locations being most myopic at the 20 degrees temporal retinal location (−0.74 D) followed by the 30 degrees temporal retinal location (−0.43 D) and the 20 degrees and 30 degrees nasal retinal locations (both −0.36 D). At near, the differences among lenses were less pronounced with accommodation playing a role in the final near profile. On average, no lenses resulted in myopic defocus when looking at a 33 cm target. Significant differences in near peripheral defocus are shown in Figure 1B. The +2.50 multifocal resulted in less hyperopic (relatively more myopic) defocus than the single vision lens at the 40 degrees and 30 degrees nasal and the 20 degrees and 30 degrees temporal retina locations and less hyperopic (relatively more myopic) defocus than the +1.50 multifocal at the 20 degrees and 30 degrees temporal retina locations. 

Figure 1.

 

Average defocus profile by retinal location averaged across the baseline, 1-year, and 2-year visits by treatment group while children looked at (A) distance and (B) a near target at 33 cm. Significant differences in peripheral defocus (P < 0.05) between each lens type at each location are noted (∗ difference between +2.50 add and single vision; † difference between +2.50 add and +1.50 add; ‡ difference between +1.50 add and single vision). Error bars represent standard error of the mean. N = nasal retina; T = temporal retina.

Aggregate Defocus Metrics, Central Defocus at Near, and Axial Growth

The mean defocus by treatment group for the aggregate metrics of defocus and central defocus at near are in Table 2. Each defocus metric evaluated differed between treatment groups (P value range = < 0.001 to 0.04) with the +2.50 add group being significantly different than the single vision group for all defocus metrics (all P ≤ 0.01). For peripheral defocus at distance, the +2.50 multifocal on average resulted in myopic defocus while the single vision lens resulted in hyperopic defocus. The most myopic meridian at distance for the +2.50 multifocal on average was more myopic than with the single vision lens. For aggregate peripheral defocus at near, all lens types resulted in hyperopic defocus; however, the +2.50 multifocal was less hyperopic (relatively more myopic) than the single vision lens. For central defocus at near, the +2.50 multifocal on average resulted in 0.21 D more hyperopic defocus than single vision lenses (P < 0.001). 

Table 2.

 

Mean (± SD) Defocus Metrics (in Diopters) by Treatment Group Evaluated in Models of the 3-Year Change in Axial Length

 

The association between each metric and the 3-year change in axial length for each model is in Table 3. Peripheral defocus at distance, most myopic meridian at distance, and peripheral defocus at near had statistically significant associations with the 3-year change in axial length (P value range = 0.006 to 0.03). Peripheral defocus at distance did the best at explaining a portion of the −0.23 mm reduction in axial growth over 3 years observed in the +2.50 multifocal group compared to the single vision group, accounting for 30% (−0.07 mm) of the observed treatment effect. The most myopic meridian at distance and peripheral defocus at near each explained only 13% and 8% of the treatment effect, respectively. Central defocus at near was not associated with the 3-year change in axial length (P = 0.56). In each model, there was no significant interaction between the defocus signal and mesopic pupil size, indicating no evidence that pupil size influenced the slowing of axial growth associated with defocus (P value range = 0.74 to 0.93). 

Table 3.

 

Association Between Defocus Metrics and 3-Year Axial Growth in Models That Do Not Include Treatment Group

 

Defocus at Distance by Retinal Location and Axial Growth

The association between peripheral defocus at each retinal location (see Fig. 1A) and the 3-year change in axial length for each model is in Table 4. Myopic defocus was associated with slower axial elongation over 3 years at all temporal retina locations (20 degrees, 30 degrees, and 40 degrees) and the 40 degrees nasal retina location (P value range = < 0.001 to 0.01). Of these locations, peripheral defocus at 30 degrees temporal did the best at accounting for a portion of the 3-year −0.23 mm reduction in axial growth in the +2.50 multifocal group, explaining 48% of the treatment effect, followed by the 20 degrees temporal retina explaining 43% of the treatment effect. 

Table 4.

 

Association Between Retinal Defocus at Specific Locations and 3-Year Axial Growth in Models That Do Not Include Treatment Group

 

There was a significant association between 20 degrees nasal retina defocus and axial growth; however, the sign of the estimate was in the opposite direction of theory (i.e. hyperopic defocus was associated with slower progression). There was no association between the 30 degrees nasal retina location and axial growth (P = 0.46). In all models of peripheral defocus by location, there were no significant interactions between defocus and mesopic pupil size, indicating no evidence that the slowing of axial growth associated with defocus at that location was enhanced by larger pupils (P value range = 0.11 to 0.98). 

Treatment Group Versus Defocus in Models

The associations between defocus with each metric or retinal location and the 3-year change in axial length when treatment group was included in the model are in Table 5. When the treatment group was added to each model allowing it to compete with defocus to explain axial growth, being in the +2.50 multifocal group was consistently associated with slower 3-year axial growth (all P < 0.001). Only two defocus metrics remained significant in the hypothesized direction with treatment group in the model. 

Table 5.

 

After Adding Treatment Group to the Model, the Remaining Association Between Defocus With Each Metric or Retinal Location and 3-Year Axial Growth

 

Each diopter of relatively more myopic (less hyperopic) aggregate peripheral defocus at near was associated with an additional −0.053 mm slowing of eye growth over 3 years, but accounted for only −0.01 mm (6%) of the observed slower axial growth versus 8% when treatment group was not in the model. Each diopter of myopic defocus on the 30 degrees temporal retina was associated with an additional −0.04 mm slowing of eye growth over 3 years, but only accounted for −0.06 mm (25%) of the slower axial growth versus 48% when the treatment group was not in the model. 

Contrary to the defocus theory, the association between defocus and axial growth was not in the hypothesized direction in two models after adding the treatment group. The significant association between hyperopic defocus on the 20 degrees nasal retina and slower growth remained. Whereas 30 degrees nasal retina defocus was not associated with axial growth without the treatment group in the model, there was a significant association after adding the treatment group but with hyperopic defocus at that location being associated with slower eye growth. No other defocus metrics were significant after adding the treatment group to the model. There was also no interaction between the treatment group and any of the aggregate defocus metrics (all P ≥ 0.47) or defocus at any location (all P ≥ 0.17), providing no evidence that the amount of defocus experienced when wearing a particular lens type further modified the effect of the treatment group. 

Pupil Size and Treatment Effect

Neither mesopic pupil size (P = 0.37, 95% CI = −0.07 to 0.19 mm axial length change/mm pupil size) nor photopic pupil size (P = 0.92, 95% CI = −0.13 to 0.15 mm axial length change/mm pupil size) modified the slowing of axial length associated with the +2.50 multifocal (Fig. 2). Mesopic (P = 0.25, 95% CI = −0.05 to 0.21 mm axial length change/mm pupil size) and photopic pupil size (P = 0.20, 95% CI = −0.05 to 0.23 mm axial length change/mm pupil size) had no effect on the change in axial length associated with the +1.50 multifocal group, and there was no overall effect of pupil size on the 3-year change in axial length (mesopic, P = 0.13, and photopic, P = 0.99). 

Figure 2.

 

Scatterplot of the 3-year change in axial length versus pupil size averaged over all visits under mesopic (left) and photopic (right) lighting conditions. The regression lines show the lack of association found between pupil size and axial growth across treatment groups (single vision [SV], +1.50 add, and +2.50 add multifocal contact lenses).

Discussion

Of the aggregate metrics of defocus that equally weighted peripheral defocus across the retina, mean peripheral defocus at distance performed the best at explaining a portion of the −0.23 mm 3-year slowing of axial growth in children wearing +2.50 multifocal contact lenses in the BLINK Study. That said, even without the treatment group in the model, this variable only explained 30% of the treatment effect. Once the treatment group was added to the model, the strongest aggregate metrics (mean peripheral defocus at distance and mean most myopic meridian at distance) were no longer significant, indicating the lens type worn better explained the slower eye growth than averaged peripheral defocus or the most myopic astigmatic signal. Although lenses that only change peripheral defocus in animal models alter axial growth even with unrestricted central vision,¹² our results do not support equally weighted peripheral defocus being the signal responsible for this altered growth. 

We also evaluated distance retinal defocus by location prior to including treatment group in the models. Defocus 30 degrees temporal on the retina explained 48% of the reduction in axial growth with the +2.50 multifocal, followed by 20 degrees temporal (43%), 40 degrees nasal (26%), and 40 degrees temporal (17%). Defocus 20 degrees and 30 degrees temporal on the retina were the two locations where, on average, the +2.50 multifocal caused the largest myopic shifts compared to single vision lenses and caused the largest amounts of myopic retinal defocus. Although the largest myopic shift was found at 40 degrees nasal, the average defocus at this location was still hyperopic (+0.23 D), which could suggest the poorer association at that location was because myopic defocus was often not achieved. Unexpectedly, the opposite association was found at 20 degrees nasal (myopic defocus was associated with faster growth), although this finding could be due to proximity to the optic nerve where eye growth characteristics and defocus influence may be different due to anatomic differences. 

Most associations between location-specific defocus and the change in axial length were no longer significant after adding treatment group to the model. Only 30 degrees temporal defocus remained significant, explaining 25% of the treatment effect versus 48% before adding treatment group. The 30 degrees temporal location did have one of the highest myopic changes in defocus with multifocal contact lenses; however, other locations either lost significance after adding treatment group to the model or had the sign of the association flip (hyperopic defocus now associated with slower axial growth). These defocus signals losing significance or having a minimized and mixed association with axial growth suggests that defocus accounts for very little of the variability in axial elongation within treatment groups. Further research is needed to explore alternate optical hypotheses to account for the slower axial growth observed with the +2.50 additional multifocal contact lens. 

Higher accommodative lag during near work has been hypothesized to influence myopia onset and progression.^41,42 Central defocus at near was not associated with the 3-year change in axial length. Although the +2.50 multifocal group had less eye growth on average, near central defocus was +0.21 D more hyperopic on average than the single vision group despite having less peripheral hyperopic defocus. Failing to find an association between central defocus at near and the 3-year change in axial length is consistent with other longitudinal studies of children.^17,43,44 

We previously reported that +2.50 multifocal lenses symmetrically reduced peripheral eye growth in the BLINK Study.¹¹ This symmetrical growth was despite myopic defocus in the periphery when wearing +2.50 multifocal contact lenses. One might interpret the associations we found between local myopic defocus in the periphery and slower axial growth (before treatment group was added to the model) as evidence for local control. However, our previous finding that inhibition of growth was greater at the fovea than in the periphery for multifocal lens wearers does not support the local control reported in young animal models.^9,10 Our results suggest that the optical mechanism behind slowed axial growth with multifocal contact lenses is either due to spatial integration across the retina or some other mechanism than defocus. 

Few longitudinal clinical trials have concurrently measured peripheral defocus with an optical correction to test whether defocus can explain observed treatment effects. The Study of Theories about Myopia Progression (STAMP) measured peripheral defocus in myopic children wearing single vision or progressive addition spectacles and found an association between myopic defocus on the superior retina caused by the progressive addition spectacle add and slower myopia progression, but the overall slowing after 1 year, although statistically significant, was clinically small (0.18 D less myopia progression).^17,19 Sankaridurg et al. evaluated relative peripheral hyperopia (the difference between peripheral and central defocus) in myopic children wearing a novel contact lens design compared to a matched control group wearing single vision spectacles and found a statistically significant association between more relative peripheral hyperopia and faster myopia progression over a 1-year period; however, the associations were small explaining very little of the variability of myopia progression over 1 year.⁴⁵ Studies have also evaluated whether dioptric changes in corneal curvature due to mid-peripheral steepening caused by orthokeratology (a surrogate for changes in peripheral defocus) are associated with changes in axial growth and have reported mixed results; one study found an association whereas another did not.^46,47 When our results are considered in context with these other studies, findings of an association between peripheral defocus and axial growth in myopic children are inconsistent and often weak when an association is present. 

Our lack of a robust association between peripheral defocus and axial growth suggests that optical attributes other than defocus could be responsible for the myopia control effect. A limitation of the Grand Seiko autorefractor used to measure defocus is that it averages over a 2.3 mm diameter⁴⁸ and does not measure higher-order aberrations. Defocus values do not capture the complex effects of both lower-order and higher-order aberrations. Peripheral aberrometry is necessary to explore alternate theories of how multifocal lenses slow eye growth involving image quality, changes in retinal contrast, depth of focus, or anisotropy.⁴⁹ Based on the analysis of a number of myopia control studies, it has also been proposed that there may be a maximum cumulative treatment effect possible over time regardless of the rate of myopia progression.⁵⁰ If the size of the treatment effect is constant regardless of the underlying hypothesized progression rate or if there is some individual threshold level of defocus necessary to slow eye growth, we would not expect a correlation between the amount of defocus and axial growth, which would create a challenge in isolating the optical factor related to the treatment effect whether that factor is defocus or some other optical signal. 

Another limitation is that due to the time involved in measuring peripheral refraction in children and contact lens decentration challenges if the eye must be rotated, peripheral measurements while wearing contact lenses were limited to the horizontal meridian and two viewing locations (distance and a 3 diopter near demand). The selected measurement conditions do not fully capture the complex dioptric visual scenes to which the eye is exposed on a daily basis.² Mountable devices that can objectively capture information regarding viewing distance and the visual environment could be beneficial in future studies. 

Due to the time required to measure peripheral refraction with the Grand Seiko autorefractor and study examination flow considerations, we measured near defocus with the habitual lens power and distance defocus with the updated contact lens power being dispensed at that visit. Because updating the contact lens power immediately before measuring near defocus would not have allowed time to adapt to the new power and could have had a temporary effect on accommodation, we chose to measure near defocus with the habitual lens worn to the visit (representing what the child had most recently experienced at near). Because distance defocus was measured after cycloplegia, we updated the contact lens power to what was being dispensed to optimize vision (capturing the maximum amount of defocus caused by the lens with the updated prescription). In reality, the eye experiences a continuum of defocus between optimal correction and what is experienced 6 months later after any myopia progression. 

We also did not find evidence that pupil size modified the magnitude of the treatment effect when children wore multifocal lenses in the BLINK Study, again suggesting that a mechanism other than peripheral defocus may be involved. Larger pupils have been hypothesized to result in a greater treatment effect by exposing more of the retina to plus power. Although a study of children fitted with orthokeratology reported that larger pupils resulted in a greater slowing of axial growth versus matched controls wearing single vision spectacles, the children with smaller pupils in their orthokeratology group also had 0.27 mm faster eye growth than children with smaller pupils wearing single vision spectacles, suggesting the pupil size effect may be confounded.³³ Another orthokeratology study that evaluated pupil size and eye growth over 2 years also reported that larger pupils were associated with slower eye growth in univariate models, although the effect of pupil size was no longer significant in a multivariate model that controlled for other factors associated with eye growth.³⁴ A study evaluating the combined effect of orthokeratology and low-dose (0.01%) atropine reported larger pupils and a better treatment effect with this combination therapy versus orthokeratology alone, but it is unclear whether the enhanced effect is due to a larger pupil with atropine or a synergistic effect via a pharmacological mechanism.⁵¹ A study evaluating the +2.50 multifocal used in the BLINK Study in combination with 0.01% atropine and matched controls from the BLINK Study did not find additional significant slowing of axial growth with adding low-dose atropine despite an increase in pupil size.⁵² 

Conclusions

Although peripheral defocus signals were found that were associated with up to nearly 50% of the slowing of axial growth observed when children wore +2.50 center-distance multifocal contact lenses, these associations were greatly diminished or no longer significant when the treatment group to which a child was assigned was added to the model. Defocus no longer being significant once treatment group was added to a model suggests that there is another factor related to the optics of the multifocal contact lens that better explains the slowing of axial growth, or that there is not a linear dose-response relationship. We also found no evidence that pupil size modified the magnitude of the +2.50 multifocal treatment effect, again suggesting that a mechanism other than peripheral defocus may be involved. These results do not support using pupil size as a criterion when deciding which children with myopia to treat with the multifocal contact lens used in the BLINK Study. Peripheral aberrometry to evaluate more complex optical signals than defocus is needed to understand the mechanism by which these lenses slow axial growth. 

Acknowledgments

BLINK Study Group 

Executive Committee: Jeffrey J. Walline, OD, PhD, FAAO (Study Chair); David A. Berntsen, OD, PhD, FAAO (UH Clinic Principal Investigator); Donald O. Mutti, OD, PhD, FAAO (OSU Clinic Principal Investigator); Lisa A. Jordan, PhD, FAAO (Data Coordinating Center Director); Donald F. Everett, MAS (NEI Program Official, 2014–2019); and Jimmy Le, ScD (NEI Program Official, 2019-present). 

Chair's Center: The Ohio State University College of Optometry; Columbus, OH. 

Kimberly J. Shaw, CCRP (Study Coordinator); Jenny Huang Jones, OD, PhD, FAAO (Investigator, 2014–2019); and Bradley E. Dougherty, OD, PhD, FAAO (Consultant, 2020-present). 

Data Coordinating Center: The Ohio State University College of Optometry; Columbus, OH. 

Loraine T. Sinnott, PhD (Biostatistician); Matthew L. Robich, MPH (Biostatistician, 2020-present); and Pamela W. Wessel (Program Coordinator, 2014–2017, deceased). 

University of Houston Clinic Site: University of Houston College of Optometry; Houston, TX. 

Laura Cardenas (Clinic Coordinator); Krystal L. Schulle, OD, FAAO (Examiner, 2014–2019); Dashaini V. Retnasothie, OD, MS, FAAO (Examiner, 2014–2015); Amber Gaume Giannoni, OD, FAAO (Examiner); Anita Tićak, OD, MS, FAAO (Examiner); Maria K. Walker, OD, PhD, FAAO (Examiner); Moriah A. Chandler, OD, FAAO (Examiner, 2016-present); Mylan T. Nguyen, OD, MS, MSPH, FAAO (Data Entry, 2016–2017); Lea A. Hair, OD, MS (Data Entry, 2017–2019); Augustine N. Nti, OD, PhD, FAAO (Data Entry, 2019–2022); Justina R. Assaad, OD, FAAO (Examiner, 2022-present); and Erin S. Tomiyama, OD, PhD, FAAO (Examiner, 2019–2022). 

Ohio State University Clinic Site: (The Ohio State University College of Optometry). 

Jill A. Myers (Clinic Coordinator); Alex D. Nixon, OD, MS, FAAO (Examiner, 2014–2019); Katherine M. Bickle, OD, PhD, FAAO (Examiner, 2014–2020); Gilbert E. Pierce, OD, PhD, FAAO (Examiner, 2014–2019, deceased); Kathleen S. Reuter, OD (Examiner, 2014–2019); Dustin J. Gardner, OD, MS, FAAO (Examiner, 2014–2016); Matthew Kowalski (Examiner, 2016–2017); Ann Morrison, OD, PhD, FAAO (Examiner, 2017–2019); Danielle J. Orr, OD, MS, FAAO (Examiner, 2018-present); Elizabeth Day, OD, MS (Examiner, 2019-present); and Rachel Fenton, OD, MS (Examiner, 2020-present). 

Data Safety and Monitoring Committee: Janet T. Holbrook (Chair, Johns Hopkins Bloomberg School of Public Health), Jane Gwiazda (Member, New England College of Optometry), John Mark Jackson (Member, Southern College of Optometry), and Charlotte E. Joslin (Member, University of Illinois at Chicago) 

Supported by National Institutes of Health (NIH) U10-EY023204, U10-EY023206, U10-EY023208, U10-EY023210, P30-EY007551, and UL1-TR002733. Bausch + Lomb (contact lens solution). 

Disclosure: D.A. Berntsen, Bausch + Lomb (F); A. Ticak, Bausch + Lomb (F); L.T. Sinnott, Bausch + Lomb (F); M.A. Chandler, Bausch + Lomb (F); J.H. Jones, Bausch + Lomb (F), Alcon Research LLC (E); A. Morrison, Bausch + Lomb (F); L.A. Jones-Jordan, Bausch + Lomb (F); J.J. Walline, Bausch + Lomb (F), Myoptechs Inc. (C); D.O. Mutti, Bausch + Lomb (F), Vyluma (C)