Comparison of the performance of myopia control in European children and adolescents with defocus incorporated multiple segments (DIMS) and highly aspherical lenslets (HAL) spectacles

Abstract

Purpose

A performance comparison of two myopia control spectacle lens designs, defocus incorporated multiple segments (DIMS) and highly aspherical lenslets (HAL), at slowing myopia progression in a European child/adolescent population. Previous research directly comparing these designs has been limited to Chinese participants and 1-year follow-up. The prevalence of myopia in European child/adolescent has been estimated at 22.60%.

Methods

Retrospective cohort study of individuals (6–17 years) with myopia progression. Participants wore DIMS (Hoya MiyoSmart) or HAL (Essilor Stellest) spectacles for a minimum of 2 years. Axial length (AL) and cycloplegic autorefraction (spherical equivalent refraction (SER)) were measured at baseline and 1 and 2 years.

Results

Mean 1-year SER changes for DIMS were −0.34D (±0.46 SD) and HAL −0.30D (±0.30); 2-year changes for DIMS were −0.50D (±0.64 SD) and HAL −0.63D (±0.56). Mean 1-year AL increases for DIMS were 0.19 mm (±0.56) and HAL 0.15 mm (±0.47); 2-year increases for DIMS were 0.29 mm (±0.63) and HAL 0.32 mm (±0.72). For equivalence margins of 0.25D and 0.50D for SER at 1 and 2 years, respectively, and similarly 0.20 mm and 0.30 mm margins for AL, DIMS and HAL lenses were equivalent apart from AL at 1 year where the 0.21 mm 95% CI upper limit just exceeded 0.20 mm. At both 1 and 2 years, none of the differences in mean SERs or ALs between DIMS and HAL were clinically or statistically significant (p≥0.05 Mann–Whitney U test). Using linear mixed model analysis, the interaction between lens type and time did not significantly affect SER or AL at 1- or 2-year follow-up (p≥0.05). 38.4% of children/adolescents with DIMS had no SER progression at 2 years, compared with 21.9% with HAL (p=0.047).

Conclusion

In a European population, DIMS and HAL lenses are essentially equivalent in their ability to reduce myopia progression and AL elongation over a 2-year follow-up period.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• DIMS and HAL spectacle lenses can both slow myopia progression in children, but research directly comparing these lens types is limited to Chinese children and to 1-year follow-up. This retrospective study directly compares DIMS and HAL in a European child/adolescent population over 2 years of spectacle wear.

WHAT THIS STUDY ADDS

• In a European population, DIMS and HAL spectacle lenses are equally effective in their ability to slow myopia progression over a 2-year follow-up period.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Based on this study, parents/carers and clinicians deciding whether to opt for DIMS or HAL lenses for myopia control can be confident that both lens types will be equally effective in a European population.

Introduction

Myopia prevalence is increasing rapidly worldwide, especially in East and South-East Asia where up to 80% of young adults are myopic, with as many as 20% of children having more than −6.00D of myopia.¹ A 2024 systematic review and meta-analysis estimated the pooled prevalence of myopia in European children and adolescents to be 22.6%.² It is estimated that in 2050, 50% of the world population¹ and 63% of the European population will be myopic.³ The rate of increase in myopia prevalence has accelerated since SARS-CoV-2.^{4 5} Myopia development and progression are influenced by genetics and environmental factors including reduced exposure to sunlight and level of educational attainment.5,7 Eyes with high myopia are more likely to develop sight-threatening conditions, including retinal degenerations and glaucoma,^{8 9} and increased prevalence of these myopia-related conditions will lead to an increasing public health and economic burden.¹⁰

This background has stimulated research into solutions that can slow myopia progression,¹¹ including the instillation of atropine eye drops,¹² dual-focus contact lenses and spectacle lenses, and orthokeratology.¹³ The burgeoning numbers of new-generation spectacle lens designs for myopia control include defocus incorporated multiple segments (DIMS) and highly aspherical lenslets (HAL). Outside a central clear zone correcting distance refractive error, both DIMS and HAL lenses provide myopic defocus in the mid-peripheral area. HAL achieves this via 11 rings of continuous lenslets which produce a volume of myopic defocus in front of the retina,¹⁴ and with DIMS, the central clear zone is surrounded by an annular peripheral zone consisting of multiple lenslets approximately 1 mm in diameter, each providing +3.5D of defocus.¹⁵ DIMS and HAL were directly compared in two 1-year retrospective studies of Chinese children,^{16 17} but have not been compared in European populations. Our aim is to compare the performance over 2 years of DIMS and HAL in slowing myopia progression, as measured by spherical equivalent refraction (SER) and axial length (AL), in a population of European children/adolescents.

Methods

Study design

The study was a retrospective cohort study set in a paediatric ophthalmology clinic with a reputation for myopia control (San Giuseppe Hospital, Milan). The clinical records database was searched for potential participants. These were patients who, between January 2021 and April 2024, had commenced myopia control treatment with either DIMS or HAL spectacles and had achieved at least 2 years wear. Patients who had previously undergone myopia control at the clinic with low-dose atropine (usually 0.01%) were eligible for inclusion, provided atropine had not been instilled for at least 6 months prior to wearing DIMS or HAL spectacles and provided other inclusion/exclusion criteria were met (table 1).

Table 1. Selection criteria.

Inclusion criteria	Children/adolescents aged 6–17 years with myopia progression European ethnicity Myopia with SER ≤−0.50D Two continuous years wear of either DIMS or HAL spectacles Attended 1- and 2-year follow-up visits
Exclusion criteria	Genetic syndromes suspected (eg, Stickler, Marfan, etc) Other eye diseases (such as glaucoma, juvenile cataracts, retinal abnormalities or any form of strabismus)

DIMSdefocus incorporated multiple segmentsHALhighly aspherical lensletsSERspherical equivalent refraction

Prior to commencing myopia control, all clinic patients (ie, not just those who participated in this study) undergo a full baseline ophthalmological assessment including symptoms and history, visual acuity (VA) with prestudy spectacles, orthoptic testing, refraction (including cycloplegic autorefraction) and dilated funduscopy. Parents/carers of suitable patients are provided with information on options for myopia control, which include DIMS or HAL spectacles. Options are discussed with patients, parents/carers and clinicians; and parents/carers are free to choose their preferred myopia control method or to opt for their children to continue wearing single-vision spectacles.

For this retrospective study, parents/carers of children wearing DIMS or HAL spectacles who had been identified as potential participants were provided with an information sheet explaining the study’s purpose, which stressed that only pseudo-anonymised data would be analysed and disseminated in an aggregated form, and which sought parental/carer consent for data analysis of their child/adolescent’s clinical data.

Outcome variable data presented were collected at the patients’ 1- and 2-year regular follow-up reviews. AL was measured using a Zeiss IOL Master, with the mean of three readings used for analysis. Refractive error was measured (Retinomax), with the mean of three readings used to calculate SER. All AL and cycloplegic autorefraction SER measurements were taken by an ophthalmologist with support from an orthoptist.

The study received approval from the Human Subjects Ethics Committee of the University of Milan and was performed in accordance with the ethical standards in the 1964 Declaration of Helsinki and its later amendments. The paper followed the STROBE checklist of items that should be included in reports of cohort studies.

Patient and public involvement

Participants/parents/carers made an informed choice as to which lens design (DIMS or HAL) was incorporated into their myopia control spectacles. Their choice of lens design determined the group (DIMS or HAL) to which they were assigned.

Interventions and sample size

Both DIMS and HAL spectacles were prescribed and dispensed according to manufacturers’ fitting guides, with participants instructed to wear the spectacles as close to all waking hours as practical (eg, not for bathing or swimming). Sample size was calculated based on an equivalence test for means using two one-sided tests within a parallel-group design. The calculation determined that a sample size of at least 70 participants per group (DIMS and HAL) would achieve 81% power to detect equivalence at a 5% significance level (α=0.05). The calculation assumes that the true difference between the group means is 0, with an SD of 0.50. The equivalence margin was set at ±0.25D.

Outcome variables

The outcome variables were the change in AL and SER. Cycloplegic autorefraction was performed after the instillation of two drops of cyclopentolate (Allergan Ciclolux 10 mg/mL) in each eye, with the second drop instilled 5 min after the first and refraction 30 min after the first instillation.

Statistical analysis

Descriptive data are reported as the mean with SD and the median with the range of values for continuous variables and as absolute and relative frequencies for categorical variables. Any differences between groups at baseline were assessed using the Mann-Whitney U test or χ² test, as appropriate.

The equivalence analyses between groups were performed for both SER and AL at 1 year and 2 years. For each comparison, and at each time point, the difference between the means of the treated groups was calculated.

The equivalence margin, which varied depending on the parameter (SER or AL) and the year of follow-up, was predefined based on clinical judgement and data from the literature^{14 15} and was used to assess equivalence.

The variance of the difference between the means was estimated using the SD and the sample size of each group. Based on these values, a 95% CI for the difference between means was calculated.

Equivalence was determined if the resulting CI was entirely within the predefined equivalence margin for that specific comparison and time point.

To analyse the progression of SER and AL over time, mixed-effects linear regression models were employed. These models assessed the effect of time (baseline, 1 year and 2 years) on SER and AL, by accounting for potential confounders including age, gender and prior use of atropine. The interaction between lens type and time was included in the models to evaluate whether the impact of lens type on SER and AL differed across the follow-up periods. Fixed effects were specified for time, lens type, age, gender and atropine use, while random effects were included to account for individual variability. Results were expressed as estimates with standard errors, and statistical significance was assessed using p values.

All analyses were performed with R V.4.1.3, and a p value <0.05 was considered statistically significant.

Results

Records were extracted from 146 myopic participants (DIMS, n=73; HAL, n=73); mean age 11.3 (± 2.36) years. Baseline participant characteristics are in table 2. As there was no statistically significant difference at baseline between right and left eye SERs (p=0.22) or right and left eye ALs (p=0.38), only right eye data were considered for inferential analysis.

Table 2. Participant characteristics at baseline.

	Total (n=146)	DIMS (n=73)	HAL (n=73)	P value
Age (years)
Mean±SD	11.3±2.36	11.2±2.3	11.4±2.38	0.89
Median (range)	11.0 (6.0–17.0)	11.0 (6.0–17.0)	11.0 (7.0–17.0)
Gender
Female	73 (50.0%)	34 (46.6%)	39 (53.4%)	0.41
Male	73 (50.0%)	39 (53.4%)	34 (46.6%)
Family history of myopia
Yes	99 (67.8%)	45 (61.6%)	54 (74.0%)	0.11
Age at diagnosis (years)
Mean±SD	7.6±2.30	7.7±2.44	7.6±2.16	0.63
Median (range)	7.5 (2.0–15.0)	8.0 (2.0–15.0)	7.0 (3.0–13.0)
Previous treatment with atropine
Yes	84 (57.5%)	54 (74.0%)	30 (41.1%)	<0.001
SER (RE)
Mean±SD		−3.4D±1.63	−3.6D±1.81	0.68
Median (range)		−3.25D (−6.50 to −0.50)	−3.25D (−9.25 to −0.05)
SER (LE)
Mean±SD		−3.5D±1.44	−3.7D±1.53	0.70
Median (range)		−3.25D (−6.75 to –1.00)	−3.25D (−8.75 to −1.00)
AL (RE)
Mean±SD		24.9 mm±0.99	25.0 mm±0.99	0.50
Median (range)		24.9 mm (23.0 to 27.6)	24.9 mm (23.0 to 27.7)
AL (LE)
Mean±SD		24.9 mm±0.95	25.0 mm±0.92	0.58
Median (range)		24.8 mm (23.1 to 27.5)	25.0 mm (23.1 to 27.6)

ALaxial lengthDIMSdefocus incorporated multiple segmentsHALhighly aspherical lensletsSERspherical equivalent refraction

Equivalence analyses results are illustrated in figure 1. Mean change in SER at 1-year follow-up for the DIMS group was −0.34D (SD=0.46), and for HAL was −0.30D (SD=0.30) (figure 1A). The difference in mean SER between groups was −0.04D (95% CI −0.17 to 0.08). Based on the equivalence margin of 0.25D and a 95% CI, DIMS and HAL lenses can be considered equivalent. Similarly, at 2-year follow-up (figure 1B), the mean change in SER for the DIMS group was −0.50D (SD=0.64), and for HAL was −0.63D (SD=0.56), with a difference in means of 0.13D (95% CI −0.07 to 0.32). Based on the equivalence margin of 0.50D and a 95% CI, DIMS and HAL lenses can be considered equivalent. At both the 1- and 2-year follow-ups (figure 2A), these differences in mean SER between DIMS and HAL are neither clinically nor statistically significant (p=0.80 and p=0.05 at 1- and 2-year follow-ups, respectively, Mann–Whitney U test).

Figure 1. Equivalence of DIMS and HAL for change in (A) SER at 1-year follow-up, (B) SER at 2-year follow-up, (C) AL at 1-year follow-up and (D) AL at 2-year follow-up. AL, axial length; DIMS, defocus incorporated multiple segments; HAL, highly aspherical lenslets; SER, spherical equivalent refraction.

Figure 2. Mean±SD changes in (A) SER and (B) AL, from baseline at 1- and 2-year follow-ups. AL, axial length; DIMS, defocus incorporated multiple segments; HAL, highly aspherical lenslets; SER, spherical equivalent refraction.

The mean increase in AL at 1-year follow-up for the DIMS group was 0.19 mm (SD=0.56) and 0.15 mm (SD=0.47) for HAL, with a difference in means of 0.04 mm (95% CI −0.13 to 0.21) (figure 1C). Based on the equivalence margin of 0.20 mm and a 95% CI, the DIMS spectacle lens is not considered equivalent to the HAL lens, with the 0.21 mm upper limit of the 95% CI just exceeding the 0.20 mm equivalence margin. At 2-year follow-up (figure 1D), the mean increase in AL for the DIMS group was 0.29 mm (SD=0.63) and 0.32 mm (SD=0.72) for HAL, with a difference in means of −0.03 mm (95% CI −0.25 to 0.19). Based on the equivalence margin of 0.30 mm and a 95% CI, the DIMS and HAL lenses can be considered equivalent. At both 1- and 2-year follow-ups (figure 2B), these differences in AL between DIMS and HAL are neither clinically nor statistically significant (p=0.34 and p=0.71 at 1- and 2-year follow-ups, respectively, Mann–Whitney U test).

Differences in myopia progression between individuals wearing DIMS and HAL spectacles are presented in table 3. At 2-year follow-up 38.4% of those wearing DIMS and 21.9% wearing HAL had no SER myopic progression, and this difference in proportions was statistically significant (p=0.047). At 1-year follow-up, 8.2% of DIMS participants had more than 1.00D of SER myopic progression compared with 0% with HAL, a difference that was statistically significant (p=0.028).

Table 3. Association between DIMS and HAL and worsening in SER/increase in AL at 1- and 2-year follow-ups (χ² test).

Myopia progression (SER or AL)	DIMS	HAL	P value
SER at 1 year (any progression)
No	32 (43.8%)	28 (38.4%)	0.61
Yes	41 (56.2%)	45 (61.6%)
SER at 2 years (any progression)
No	28 (38.4%)	16 (21.9%)	0.047
Yes	45 (61.6%)	57 (78.1%)
SER at 1 year (>−0.50D progression)
No	58 (79.5%)	57 (78.1%)	0.99
Yes	15 (20.5%)	16 (21.9%)
SER at 2 years (>−0.50D progression)
No	49 (67.1%)	41 (56.2%)	0.23
Yes	24 (32.9%)	32 (43.8%)
SER at 1 year (>−1.00D progression)
No	67 (91.8%)	73 (100.0%)	0.028
Yes	6 (8.2%)	0 (0.0%)
SER at 2 years (>−1.00D progression)
No	61 (83.6%)	58 (79.5%)	0.67
Yes	12 (16.4%)	15 (20.5%)
AL at 1 year (any increase)
No	15 (20.5%)	13 (17.8%)	0.67
Yes	58 (79.5%)	60 (82.2%)
AL at 2 years (any increase)
No	11 (15.1%)	7 (9.6%)	0.31
Yes	62 (84.9%)	66 (90.4%)
AL increase at 1 year
≤0.2 mm	50 (68.5%)	58 (79.5%)	0.13
>0.2 mm	23 (31.5%)	15 (20.5%)
AL increase at 2 years
≤0.3 mm	47 (64.4%)	47 (64.4%)	0.99
>0.3 mm	26 (35.6%)	26 (35.6%)

ALaxial lengthDIMSdefocus incorporated multiple segmentsHALhighly aspherical lensletsSERspherical equivalent refraction

Analysis over time for SER and AL

The interaction between lens design and time did not significantly affect SER or AL at either 1- or 2-year follow-up. For SER: year 1 versus baseline: estimate=−0.04, SE=0.07, p=0.52; year 2 versus baseline: estimate=0.13, SE=0.07, p=0.07. For AL: year 1 versus baseline: estimate=0.04, SE=0.08, p=0.64; year 2 versus baseline: estimate=−0.03, SE=0.08, p=0.70. For both SER and AL, no significant differences were found with age (SER: p=0.22, AL: p=0.59) or with previous use of atropine (SER: p=0.22, AL: p=0.25). No significant differences were found between males and females for SER (p=0.17) but males had significantly longer AL compared with females (estimate=0.32, SE=0.14, p=0.025).

Adverse events reported during the study

Minor visual difficulties were reported by 6.8% (10/146) participants, 5.5% (4/73) DIMS and 8.2% (6/73) HAL. Some were non-specific visual disturbances, though blurred vision was also noted. Most cases resolved spontaneously within a few days; two cases (one DIMS and one HAL) lasted up to 1 month.

Discussion

The study is novel in comparing DIMS and HAL spectacle lenses for myopia control in a European population of children/adolescents over a 2-year period, using tests for equivalence not previously used to compare methods of myopia control. Our findings indicate that DIMS and HAL are essentially equivalent at reducing myopia progression, measured by change in SER and AL, over a 2-year period. There was one minor departure from equivalence: for AL at 1 year follow-up, the difference between DIMS and HAL just crossed the upper boundary of the 95% CI.

Two previous retrospective studies have directly compared DIMS and HAL in Chinese children over a 1-year period.^{16 17} Guo et al found HAL lenses more effective than DIMS at reducing myopia progression after 1 year, with significant differences between lens types for SER and AL.¹⁶ In contrast, in the current study, at both 1- and 2-year follow-ups, differences in SER and AL between DIMS and HAL were neither clinically nor statistically significant. There are notable differences between these studies: the Chinese study had markedly different sample sizes in DIMS and HAL groups; there were statistically significant differences in mean baseline age between groups; the average follow-up period of approximately 9 months necessitated the use of a correction factor to predict the 1-year SER and AL outcomes; there was much missing data; and cycloplegic refraction was not routinely performed at follow-up. All these issues were addressed in the current study, where the 1-year progression of SER with DIMS was −0.34D (SE=0.05) and for AL was 0.19 mm (SE=0.07). The equivalent 1-year progression data for DIMS in Guo et al reported greater progression, with SER progressing by −0.63D (SE=0.08) and AL increasing by 0.27 mm (SE=0.04).¹⁶ Yang et al¹⁷ used similar methods to our study: comparing DIMS and HAL alongside orthokeratology and myopia control soft contact lenses.¹⁷ Their 1-year progression findings for median SER were −0.25D for both DIMS and HAL, while 1-year mean AL increases were 0.20 mm±0.18 for DIMS and 0.19 mm±0.19 for HAL, with no statistically significant differences between DIMS and HAL for SER or AL. These results suggest both lens designs are equally effective over 1 year for myopia control; findings which are in line with the 1-year results from our study.

The double-masked DIMS randomised controlled trial¹⁵ reported lower 1-year progression rates than the retrospective studies: −0.17D (SE=0.05) for SER and 0.11 mm (SE=0.02) for AL. For 2-year data, the RCT DIMS progression rates were −0.38D (SE=0.06) and 0.21 mm (SE=0.02), again lower than for our study, which found −0.50D (SE=0.07) for SER and 0.29 mm (SE=0.07) for AL. This RCT was conducted between 2014 and 2017, pre-COVID-19 pandemic and its associated increase in myopia prevalence;^{18 19} increases mainly reported in children from Asian populations, but also observed in Spanish children.²⁰ Other factors contributing to these variations include different environmental myopic triggers²¹ encountered in Asian and European populations. The double-masked 1-year HAL RCT,¹⁴ which commenced (July 2018) pre-COVID-19 pandemic, also found lower 1-year progression rates for AL (0.13 mm, SE=0.02) than the retrospective studies. Progression rates for SER (−0.27D, SE=0.06) were lower than both our study and Guo et al. Another possible explanation for lower progression rates found in RCTs compared with the retrospective studies stems from the methodological constraints inherent in RCTs. To reduce potential confounding factors, RCTs frequently recruit very selective populations; perform more regular assessments than in standard clinical practice and use specialised clinics, settings which may differ markedly from those where standard clinical practice is conducted. Therefore, RCTs may not accurately describe what happens in the real world, possibly leading to overestimates of treatment effects. Real-life studies may have greater generalisability alongside their lower internal validity.²² Although the RCTs’ pre-COVID dates are likely to contribute to their lower progression rates, these RCTs may overestimate the myopia control abilities of DIMS and HAL spectacles when prescribed in real-life clinical settings. While no study design has the RCT’s scientific rigour, there is a case for using a range of methodologies;²³ and complementary use of RCTs and real-life studies can combine to create a powerful evidence-based approach to clinical research.²⁴

For individual subjects, 32/73 (43.8%) wearing DIMS had no SER myopia progression at 1 year compared with 28/73 (38.4%) wearing HAL: by 2 years these percentages were 38.4% for DIMS and 21.9% for HAL and this difference in proportions was statistically significant. For both lens designs these proportions with no increase in SER were greater than in the respective RCTs: the SER 1-year percentage from the HAL RCT was 28% and for the DIMS RCT the SER 2-year percentage was 21.5%. When comparing proportions of participants with >−0.50D or >−1.00D increases in SER, the differences between DIMS and HAL were not statistically significant apart from SER >−1.00D at 1-year follow-up where 6/73 (8.2%) of DIMS participants had >−1.00D progression compared with 0/73 (0%) with HAL. By 2-year follow-up, the relative performance of the two groups had reversed; 16.4% of the DIMS group had >−1.00D progression compared with 20.5% in the HAL group. For AL, there were no significant differences between DIMS and HAL at 1-year or 2-year follow-up for any increase in AL.

Strengths and limitations

Most DIMS and HAL studies, including those directly comparing DIMS and HAL, have been in Asian populations, and our study is an important extension of earlier research on a European population. Another strength is the novel inclusion of the equivalence analysis.

One weakness is the potential source of bias introduced by the absence of random allocation to the intervention groups (DIMS or HAL). Indeed, the selection of intervention was an informed decision taken by participants/parents/carers themselves. It can be argued that the undoubted benefits of random allocation in terms of reducing bias are somewhat offset by random allocation being remote from the reality of the patient selection process in clinical practice. Random allocation and the other eligibility criteria applied in RCTs can make results less relevant to clinical practice, leading to the suggestion that a diversity of research methodologies, including non-randomised observational studies, should be adopted.²³

One selection criterion for this study was that the child/adolescent had attended both 1- and 2-year follow-up visits, so there were no missing data. This is a strength as no imputations for missing data were required but also a possible weakness limiting the generalisability of the results to less-compliant populations. Similarly, results from this European population may not be generalisable to Asian or other populations.

Almost 75% of the DIMS group and over 40% of the HAL group had undergone previous low-dose atropine (usually 0.01%) treatment at the clinic, and this difference in proportions was statistically significant. To counteract this possible weakness, participants were required to have ceased atropine instillation at least 6 months before commencing wear of DIMS or HAL spectacles. The ATOM2 study established that after a 2-month washout period following 2 years of myopia control with 0.01% atropine, photopic pupil diameters had returned to baseline levels; distance and near logMAR VAs were improved compared with baseline; and accommodation was lower than baseline by an amount consistent with the greater than 2-year duration since the baseline measurements.²⁵ Therefore, the likelihood of previous use of low-dose atropine affecting our study results is remote and this was confirmed by the linear mixed model analysis which showed that previous atropine use was not a factor significantly affecting results.

Conclusion

In a European population, DIMS and HAL spectacle lenses are essentially equivalent in their ability to reduce myopia progression and AL elongation over a 2-year follow-up period.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.