This meta-analysis pooled 12 studies covering 834,814 intravitreal anti-VEGF injections and 119 retinal detachment events, to put a number on a complication that patients ask about and clinicians usually quote from memory.
The pooled risk of rhegmatogenous detachment was 0.012% per injection, about 1 in 8,675, with low heterogeneity. Restricting to events within 90 days gave essentially the same figure. Subgroup analyses found no difference by injection volume, single versus multi-centre setting, indication, or who gave the injection.
The more useful number is the per-eye one: 0.08%, about 1 in 1,250 eyes, with a much wider interval. That is the figure a patient starting long-term treatment actually needs, and the authors point out that the mean of 13.8 injections per eye in these studies probably understates real cumulative exposure — so 1 in 1,250 is likely optimistic.
One subgroup stands apart. A single study of severe proliferative diabetic retinopathy reported tractional detachment at 3.6%, roughly 1 in 28, which almost certainly reflects fibrovascular contraction after VEGF withdrawal rather than the injection itself. That is a different complication with a different mechanism, and it should not be folded into the reassuring number.
- Quote per-eye risk, not per-injection, when consenting someone for a course of treatment.
- About 1 in 1,250 eyes is the usable figure, and it likely understates long-term exposure.
- Severe proliferative diabetic retinopathy is a separate conversation: tractional detachment risk there was roughly 1 in 28.
- No difference by operator or setting, so this is not a technique or volume argument.
- Certainty was low; treat these as best current estimates rather than settled numbers.
The statistics, in plain English
Per-injection and per-eye risk answer different questions and differ here by roughly sevenfold. A patient does not receive one injection; they receive a course, so the per-eye denominator is the one that matches what they are consenting to. Notice the confidence intervals: the per-injection estimate is tight, 0.009 to 0.015%, while the per-eye estimate runs 0.02 to 0.31% — an order of magnitude of uncertainty, because far fewer studies reported eye-level denominators. I-squared of 14.4% on the main estimate means the studies agreed well. The 3.6% tractional figure comes from one study and should be read as an alert, not an estimate.
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.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for ophthalmology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free