HAL lenses outside Asia: the US trial, and why 71% and 53% describe the same result

Jagrut Lallu — BOptom (Hons), MSc Specialty Lenses (Hons), FIAOMC. Therapeutic Optometrist & Contact Lens Specialist. Founder, New Zealand Eye Research Centre; Partner, Rose Optometry; WCO Asia-Pacific Myopia Ambassador; International Myopia Institute NZ Ambassador; Clinical Senior Lecturer, Deakin School of Optometry; Honorary Teaching Fellow, University of Auckland.
Much of the early evidence for spectacle lenses with highly aspherical lenslets (HAL, sold as Essilor Stellest) came from East Asia. A randomised trial across nine US sites, published in JAMA Ophthalmology in 2026, tests the design in a more ethnically mixed group of children. Its two headline figures, 71% and 53%, sound like different results. They are not.
How the trial was run
This was a double-masked, parallel-group randomised clinical trial. In total, 159 children aged 6 to 12 years with manifest spherical equivalent refraction between −0.75 and −4.50 D were randomised 1:1 to HAL lenses (n = 77) or standard single-vision spectacle lenses (SVL, n = 82) of the same material. Reported wear averaged more than 12 hours a day in both groups.
There were two coprimary outcomes, both of which had to show a benefit: change in cycloplegic spherical equivalent refraction (cSER) and change in axial length (AL), the physical length of the eye. A masked technician took the measurements.
One detail matters. The trial's primary end point is at 36 months. This paper is a prespecified 24-month interim analysis carried out "for regulatory purposes", supporting US FDA authorisation of the device. Three-year results have not yet been reported.
The between-group results at 24 months
Over two years, refraction changed by −0.25 D with HAL and −0.90 D with SVL, a difference of 0.64 D (95% CI, 0.50 to 0.79; P < .001). Axial length grew by 0.21 mm with HAL and 0.45 mm with SVL, a difference of −0.24 mm (95% CI, −0.29 to −0.19; P < .001).
A 95% confidence interval (CI) is the range of true effects reasonably compatible with the data; here even the least favourable end of each interval favours HAL. The P value describes how surprising a difference this large would be if the lenses made no difference. Both cleared the stricter interim threshold (α = .009).
At the eye level, 75.0% of HAL eyes (102/136) progressed by less than 0.50 D, compared with 36.2% of SVL eyes (47/130). For axial length, 56.6% of HAL eyes (77/136) grew by less than 0.20 mm, compared with 15.9% of SVL eyes (21/132). The Discussion quotes a different HAL axial figure; we have used the Results.
Why 71% and 53% describe the same result
The authors report a 71% relative reduction in refraction change and a 53% relative reduction in axial elongation, from the same children over the same two years. They differ for two reasons.
First, relative figures depend on the control group. Each percentage is the between-group difference expressed as a share of the SVL group's change (−0.90 D and 0.45 mm). Had control children progressed more slowly, the same absolute benefit would give a smaller percentage.
Second, refraction and axial length measure different things. Axial length reflects the eye growing longer, which drives myopia and its long-term risks. Refraction combines axial length with the focusing power of the cornea and lens, and depends on how fully focusing is relaxed during testing; here cycloplegia used tropicamide 1% alone, which is weaker than cyclopentolate. Eyes also lengthen naturally with growth: the authors estimated that eyes with stable refraction still elongated by 0.12 mm (HAL) and 0.21 mm (SVL) over two years. That expected growth sits in the denominator of the axial percentage, which is one reason it is smaller.
For families, the absolute numbers are the more useful anchor: on average, 0.64 D less refraction change and 0.24 mm less axial growth over two years than with single-vision lenses.
Subgroups and safety
Age subgroups showed refraction differences of 1.03 D (6–8 years), 0.54 D (9–10 years) and 0.55 D (11–12 years). The trial was not powered to detect subgroup differences and no interaction test was significant, so these figures are descriptive only.
Distance best-corrected acuity at 24 months was −0.04 logMAR in both groups, and no child lost more than one line. The Results report 21 ocular adverse events (9 HAL, 12 SVL) in 11 participants, none considered related to the test lens; one control-group headache was recorded as possibly device related. The abstract presents 9 and 12 as participant counts, a minor inconsistency.
Limitations of this evidence
- Sponsor authorship and analysis. Essilor International funded the trial and was involved in its design, conduct, analysis, interpretation and writing. Two authors are Essilor employees and two work for the contract research organisation that ran it. Statistical analysis was done by Essilor and contract research organisation authors; no independent outcome analysis is described.
- Interim, not final. These are 24-month data from a 36-month trial; longer-term and post-treatment data are still needed.
- Uneven dropout. 9.1% of HAL children discontinued versus 19.5% of SVL children. Multiple imputation and tipping-point analyses suggest the result is robust, but the imbalance is worth noting, and retention figures quoted in the Discussion are not reconciled with these.
- Surrogate outcomes. The authors acknowledge that the clinical relevance of the effect into adulthood, including any change in the risk of myopia-related disease, is unknown.
- Who was included. Children with more than −4.50 D of myopia or 1.50 D of astigmatism were excluded. South Asian children numbered 2 (HAL) and 3 (SVL), and only 1 Native Hawaiian or Other Pacific Islander child took part, so the trial says little directly about Māori, Pacific or South Asian children in New Zealand.
- Wear and masking. Wear time was self- or parent-reported, and someone trained to spot the lenslet rings could be unmasked.
- Comparator. HAL was compared with single-vision lenses, not another active treatment.
What this means for families in Hamilton
This trial adds evidence from a mixed, non-Asian population that HAL lenses may slow myopia progression on average compared with standard spectacles, at least over two years. It cannot predict an individual child's response, and the effect needs confirming at three years. Our reviews of the two-year DIMS lens versus 0.01% atropine trial and an orthokeratology trial using a smaller treatment zone show how other approaches have been tested.
The right approach depends on a child's age, prescription, eye measurements and lifestyle, and should follow a professional assessment. Our optometrists, including Jagrut Lallu and Jacqueline Rowe, provide myopia control at Rose Optometry in Hamilton. To talk through the evidence first, book a free Myopia Chat with Jagrut.
References
- Shen J, et al. Spectacle Lenses With Highly Aspherical Lenslets for Myopia Control: A Randomized Clinical Trial. JAMA Ophthalmology. 2026;144(10):889-896. PMID: 42623028. DOI: 10.1001/jamaophthalmol.2026.3288





