The UK's licensed atropine dose ranks 16th of 24, and stops cleanest

Rank 23 childhood myopia treatments and an untreated control on a year's refraction change, and the dose licensed in the UK comes 16th. The analysis, published in Ophthalmology and Therapy on 7 September, pooled 95 randomised trials of atropine for myopia control and its optical and light-based rivals. Earlier syntheses ranked on treatment alone, the authors say. This one ranks on what happens 12 months after it stops too, and the order turns over.

That matters here. Ryjunea, atropine 0.1 mg/ml, is the first low-dose atropine the MHRA has licensed: children of 3 to 14, myopia between −0.50 DS and −6.00 DS, progressing at 0.50 DS a year or more. That is 0.01%, the bottom of the band this paper supports, and it is private-only: the NICE appraisal stalled on 11 August 2026 when Santen asked for more time on the draft guidance.

What the study did

Jovanović and colleagues found 117 randomised trials of 24 interventions in children under 18 with refraction worse than −0.50 D. Ninety-five went into the synthesis: 69 in the one-year refraction network, 80 in the axial-length network, eight in the post-cessation ones. Seventy-eight of the 117, 68 per cent, were run in East and Southeast Asia; eight were at low risk of bias throughout.

It was funded by Santen, whose drop is the licensed one: two of the seven authors are Santen employees, five IQVIA staff under contract.

What was found on atropine for myopia control

Rebound was extra progression in the 12 months after stopping, against untreated controls; over 0.25 D or 0.10 mm counted as meaningful. Stopping high-dose atropine (0.5% and above) cost −0.71 D (−1.13 to −0.31) and 0.20 mm of axial length, moderate-dose 0.19 mm. Low-dose atropine cost −0.10 D (−0.35 to 0.14) and 0.06 mm, multifocal soft lenses 0.01 mm, the intervals crossing zero.

After LDA discontinuation, progression was statistically and clinically indistinguishable from control for both SER and AL, indicating no rebound on either SER or AL.

In the dose-response model, which assumes higher doses do at least as much as lower ones, the gradient holds: −0.20 D a year for 0.05%, −0.09 D for 0.025%, none for 0.01%. And not only atropine: repeated low-intensity red light, the most effective on axial length at one year, rebounded 0.14 mm (0.02 to 0.26) once stopped, on one trial at high risk of bias and with no evidence of effect beyond year 1.

Effect during treatment runs the other way. In that same model, atropine split by concentration, 0.025% ranked 10th of the 24 and 0.01% 16th on refraction at one year, 0.025% beating it by 0.1 D (0.02 to 0.22). The licensed dose is the gentlest of the band, not the strongest, and the licence covers that strength alone: anything above sits outside it.

What the spectacle lens trials add

A two-year randomised trial in Ophthalmology Science, Zeiss-funded, followed 234 children aged 6 to 13 in Spain and Portugal, 208 to 24 months. Axial elongation was 0.21 mm in the CARE lens against 0.43 mm in single vision, and the split matters: 0.14 mm of benefit in year one, 0.08 mm in year two.

A second, non-randomised trial of 504 children in Chengdu compared orthokeratology with two designs of one lens. In moderate myopes, ortho-k and the 7 mm design gave 0.16 mm and 0.19 mm of axial growth against 0.25 mm for the 9 mm design; in low myopes nothing separated them. Only the abstract could be read, so nothing beyond it is claimed.

Why it matters in practice

Choose the exit at the same appointment as the entry. If a family starts something that rebounds, the stopping plan belongs in the consent conversation and the notes, not at 16. On 0.01% the honest framing is the smallest effect of the low doses and the cleanest stop.

Then measure axial length. The authors warn that corneal reshaping confounds refraction, and that axial length is the more informative measure for a cornea-modifying treatment. Which says to me that a practice fitting ortho-k without a biometer cannot audit its outcomes: the same case for equipment that Denmark's high-street OCT data made for imaging.

And say what is not known. Combining 0.01% atropine with orthokeratology was among the three most effective at one year and beat low-dose atropine alone over two years, with no post-cessation data. The College of Optometrists says combined therapy "is based on a limited number of trials with a low level of certainty".

The caveats

• Funded by Santen, whose 0.01% drop is the licensed UK product. It calls its results indirect comparisons resting on modelling assumptions, not prescribing guidance.
• The rebound networks rest on eight trials, more than one only for low-dose and high-dose atropine; the red light and contact lens figures on one each.
• Confidence in the one-year refraction estimates was mostly low or very low. Chinese regional databases were not searched, safety not pooled, cost out of scope.
• The lens trial lost 14 children from the single-vision arm against 4 from the lens arm in year two. The ortho-k comparison was non-randomised, single-centre and abstract-only.

Sources

1. 1.Low-Dose Atropine and Other Interventions for Pediatric Myopia: A Systematic Review and Bayesian Network Meta-analysis of Efficacy and Post-treatment Outcomes · Ophthalmology and Therapy, 7 September 2026
2. 2.Two-Year Efficacy of Cylindrical Annular Refractive Elements Spectacle Lenses for Myopia Control in European Children · Ophthalmology Science, 1 July 2026
3. 3.Effects of orthokeratology versus cylindrical annular refractive elements spectacle lenses on myopia control and choroidal thickness in myopic children: A 1-year prospective clinical trial · Contact Lens and Anterior Eye, 10 September 2026
4. 4.MHRA approves first licensed low-dose atropine treatment for childhood myopia in the UK · College of Optometrists, 10 November 2025
5. 5.Low-dose atropine eye drops for treating myopia in people 3 to 14 years [ID6517] · NICE, 11 August 2026