When an injectable blocks a blood vessel — the rarest and heaviest complication

If the tip of a needle or cannula happens to sit inside an artery when material is pushed, that material travels with or against the flow and lodges in a narrower vessel. If the blocked branch supplies skin, that area blanches and can go on to die. Some arteries at the front of the face connect with the ophthalmic artery, which supplies the eye, so with enough pressure material can be driven towards the eye.

The embolic event can be immediate, but it can also appear later, when a deposit fragments or migrates. A 2026 review divides presentations into fulminant, immediate and delayed, with different prognoses attached to each.

A 2022 review of the world literature identified 233 cases of filler-induced visual loss, of which 172 patients were left with severe visual impairment in at least one eye. The typical patient was a young woman; the material was hyaluronic acid or autologous fat; the site was the glabella or the nose.

A 2024 analysis restricted to cases since 2018 counted 379. At presentation, 60.9% had no light perception; ptosis was present in 54.3% and ophthalmoplegia in 42.7%. Mean age was 38.

For a figure with a denominator, nasal filler gives the best data. A 2024 systematic review pooling 9,657 patients who had non-surgical rhinoplasty found an overall complication rate of 39.11%, mostly erythema and swelling — but serious vascular complications such as blindness, skin necrosis and stroke in 0.27%. Two or three people in a thousand.

Rare and safe are different words. For the individual, 0.27% is all or nothing.

A 2019 review of 50 reports of visual compromise found the highest-risk procedure was nasal augmentation, followed by glabellar lines. The glabella used to be the commonest site; it has shifted to the nose, and in the last three years of that search 35% of reports involved the nose. By contrast, there were no reports of blindness from the temple or chin, and relatively few from the forehead.

One dataset on the glabella is worth knowing. A dermatologist reviewed ten years of records covering 719 glabellar injections placed with a superficial serial microdroplet technique and found no vascular occlusions. The paper's own conclusion, however, was that even with a safer technique most injectors should avoid this area. Technique can lower the risk; it does not remove it.

This comes up often in consultations and it is not true. A 2025 systematic review looked only at biostimulator injectables. What it found:

Two things have to be read together here. One is that boosters are not exempt. The other is that a difference in the number of reports is not a ranking of safety — usage volumes differ and so do reporting habits. That PCL has no reported blindness is not a reason to relax; it is a statement that none has been reported yet.

There is one difference that matters a great deal. Anything that is not hyaluronic acid cannot be dissolved with hyaluronidase. The means of reversal simply does not exist. We set this out separately in can every filler be dissolved.

An occlusion affecting skin usually declares itself in this order.

• Pain during or straight after the injection that is out of proportion — clearly worse than the rest of the treatment, or in a different place
• Blanching — the area that has lost its blood supply turns pale
• A mottled, net-like pattern — appearing within minutes to hours
• Darkening, then blisters and crusting — by this stage it is late

The eye is different. The typical picture is sudden loss of vision within minutes of injection with severe ocular pain. A drooping lid or restricted eye movement accompanies it about half the time. This looks nothing like a bruise, and it is measured in minutes, not hours.

Rarely, material reaches the brain. A 2026 review reports neurological complications in up to 24% of published cases, most often in the middle cerebral artery territory.

This is the most important section on the page, and there is little good news in it.

For a hyaluronic acid filler, hyaluronidase is the only intervention consistently associated with better outcomes. How it is delivered changes the result. A 2022 systematic review of 144 cases of visual loss found that hyaluronidase given subcutaneously and intra-arterially helped visual recovery to varying degrees, while retrobulbar injection seemed less helpful.

On retrobulbar injection there is more direct evidence. Four patients with vision loss were given one or two retrobulbar injections of 1,500 to 3,000 units. In none of the four did the retinal artery reopen, and in none did vision improve. All were treated at least four hours after onset.

Which vessel is blocked also decides the outcome. Ophthalmic artery occlusion and central retinal artery occlusion carry a very poor prognosis; branch retinal artery occlusion is the more favourable pattern. In the 379-case analysis, hyaluronidase dose showed no statistical link to recovery — the injection site did.

So: time is everything, and even time is not a guarantee. A 2014 review put it as “mostly irreversible”, adding that partial recovery has been reported with hyaluronic acid and CaHA. The 2019 review is blunter — there are no proven preventive measures or treatments for blindness after filler injection.

We do not promise that this will not happen. We prepare on two fronts instead: lowering the probability, and noticing it quickly if it occurs.

• We start from the site. The nose and glabella are where the reports concentrate. Where the same concern can be approached another way, we look at that first
• Whether the material is reversible goes into the plan. Hyaluronic acid can be dissolved; PCL, PLLA and CaHA cannot. The riskier the site, the more that difference weighs
• We treat pain during injection as information. The question is not “can you bear it” but “is this different from a moment ago”
• We say on the day what to watch for. The early signs above are that list
• We keep the means to act immediately. Hyaluronidase is kept on hand, and any eye symptom goes to an ophthalmologist without delay — because it is not something we can resolve ourselves

Which treatment to choose depends on skin condition, how much change is wanted, the site, and what has been done before. We put this complication on the table at the point where the site is decided, not afterwards.

• We found no Korean incidence data. The figures above are pooled from world literature, and how many cases go unreported is unknown. The 2014 review notes that the actual incidence may be higher than reported
• No trial has compared products against each other for this risk. The differences in the table above carry differences in usage volume and reporting habits as well
• Whether a cannula is safer than a needle is beyond this page; we cover it in needle or cannula
• The value of aspirating before injecting is disputed in the literature. It is not a reason not to do it, but doing it does not establish safety

This column was written and reviewed by Lee Chi-Hak, MD, medical director of Miso Clinic in Daegu, South Korea. It is general information and does not replace diagnosis or treatment. The same symptom can have different causes and therefore different answers.