No signal that anti-VEGF harms neurodevelopment in retinopathy of prematurity - and no study powered to find one

Design

American Academy of Ophthalmology graded evidence review, PubMed search to December 2025

Population

26 studies of infants treated for retinopathy of prematurity 2006-2022 across 39 countries

Primary outcome

Neurodevelopmental outcomes after anti-VEGF versus laser photocoagulation as primary treatment

Effect

20 of 26 studies, including all three randomised trials, found no significant difference; 4 favoured laser, 2 favoured anti-VEGF; no difference in IQ in the five studies testing beyond age four

An American Academy of Ophthalmology assessment reviewed the evidence on neurodevelopmental outcomes after intravitreal anti-VEGF compared with laser photocoagulation as primary treatment for retinopathy of prematurity. Of 52 articles retrieved, 26 met criteria: four papers on three randomised controlled trials rated level II, and 22 comparative cohorts rated level III, covering infants treated between 2006 and 2022 across 39 countries.

Twenty of the 26 studies, including every randomised trial, found no statistically significant difference in neurodevelopmental outcomes between the two treatments. Four found worse outcomes after intravitreal bevacizumab and two found worse outcomes after laser. Only five studies tested intelligence quotient beyond four years of age, and none of those found a difference. Half the studies used the Bayley Scales, and 54 per cent assessed infants before 24 months of age.

The conclusion has to be stated precisely, because it is easy to overstate in both directions. There is no level I evidence, and no study was powered for neurodevelopmental outcomes or designed to detect small differences - so this is absence of evidence of harm, not evidence of absence. What it does mean is that the neurodevelopmental worry, which has shaped many units' preference for laser, is not supported by the accumulated data, and should no longer be the deciding argument when anatomical and practical considerations favour anti-VEGF. Where it stays relevant is in consent: parents asking about the brain deserve this answer with its limitation attached, and in follow-up, because most of the assessments so far stopped before the age at which cognition can be measured well.

• Stop treating neurodevelopmental risk as the argument that settles laser versus anti-VEGF; decide on zone, stage and follow-up feasibility.
• Tell parents plainly that no difference has been found and that no study was powered to find a small one.
• Arrange neurodevelopmental follow-up beyond two years wherever it can be organised - most of the evidence stops too early.
• Remember anti-VEGF's real trade-off is the need for prolonged retinal follow-up for late reactivation, which is a service question in India.
• Record which agent and dose were used; the four studies suggesting harm all involved bevacizumab.

The statistics, in plain English

This is a graded evidence review, not a pooled analysis, so there is no summary effect size to quote - the finding is a count of studies and the level of evidence behind them. Level II and III evidence with no study powered for the outcome means a small true difference could sit undetected under all 26 studies. The six discordant studies, four against bevacizumab and two against laser, are what unpowered heterogeneous data look like when there is no real effect, but they are equally consistent with a modest one.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.