Is a Cannula Safer? Weighing Blunt Cannula Against Sharp Needle for Vascular Occlusion Risk

When patients ask me about hyaluronic acid, the first question is usually some version of "could this blind me?" Once I have walked them through the risk, the second question is almost always the same one: "So if we use a cannula, I'm safe?"

It is a good question, and it deserves a longer answer than yes or no. Cannulas do make a difference, and the difference is not small. But between "makes a difference" and "you can relax" sit several things patients are rarely told.

The difference is in the tip

A sharp needle has a bevelled, cutting tip. It gets through tissue by slicing, so resistance is low and placement is precise. For very superficial work, or for putting a small amount at one specific point, it is the handier tool.

A cannula, properly a microcannula, has a rounded, blunt tip with the opening on the side. It does not cut its way in. It pushes and glides along a tissue plane, and when it meets something elastic like a vessel or a nerve, it tends to shove it aside rather than go through it. Because the tip is blunt, a cannula cannot break skin on its own, so you make an entry port with a sharp needle first. From that one port you can then fan out in several directions.

How occlusion actually happens, and how much the device matters

Vascular occlusion is not simply a case of injecting so much that the vessel gets squeezed. The dangerous version is filler entering the lumen of an artery, travelling with the bloodstream, and lodging downstream. Facial arteries anastomose with each other and share flow with the eye, which is why trouble at the glabella or the nose can reach the retina and cause irreversible visual damage. I have written separately about the occlusion mechanism and the rescue window from blanching to necrosis.

What the device changes is the first link in that chain: whether the tip gets into the lumen at all. A sharp edge can pass straight through a vessel wall; a blunt tip tends to displace it. So mechanistically the device should matter. The question is by how much.

What the numbers say

The largest dataset on this question is the retrospective cohort by Alam and colleagues in JAMA Dermatology (2021;157(2):174-180, PMID: 33377939, DOI 10.1001/jamadermatol.2020.5102). Three hundred and seventy board-certified dermatologists reported roughly 1.7 million syringes injected. The result: one occlusion per 6,410 one-millilitre syringes with needles, and one per 40,882 with cannulas (P < .001). On multivariate analysis, cannula injections carried 77.1% lower odds of occlusion than needle injections.

That is an order of magnitude, not noise. But the same dataset holds two more figures patients seldom hear. Injectors with more than five years of practice had 70.7% lower odds of occlusion than less experienced ones. And of all reported occlusions, 85% left no long-term sequelae.

Key point: The cannula pushes risk down by an order of magnitude, yet in the very same dataset the effect of "how long the injector has been doing this" is nearly as large. The device is one variable, not the answer.

A 2025 review strikes the same tone (JPRAS Open 2025;48:415-424, PMID: 41541903, DOI 10.1016/j.jpra.2025.12.002). Having worked through twenty-five years of literature, the authors consider cannulas the safer option in high-risk, vessel-dense regions, but they tie that conclusion to two conditions: proper technique and thorough anatomical knowledge. They also say plainly that cannulas are not inherently risk-free, and that arterial wall penetration and ischaemic events remain possible.

A cannula is not insurance: two things that get misread

The first is the aspiration test. Plenty of people believe that drawing back before injecting, and seeing no blood, means the tip is not in a vessel. After reviewing the literature on this, Goodman and colleagues concluded that aspiration cannot be relied on and should not be employed as a safety measure (Aesthetic Surgery Journal 2022;42(1):89-101, PMID: 33512439, DOI 10.1093/asj/sjab036). Their reasoning is practical: a positive aspiration merely defers the risk to the next injection site, where you would again be trusting a negative result. They argue for shifting the emphasis elsewhere, namely to injection anatomy, control of plane, and slow, low-pressure delivery.

The second is that cannulas have their own reported occlusions. A 2025 case report in the Journal of Cosmetic Dermatology (2025;24(7):e70314, DOI 10.1111/jocd.70314) describes a 36-year-old woman having glabellar filler. The operator first used a 22G sharp needle to create a preformed subdermal tunnel, then switched to a 22G dull needle to inject cross-linked hyaluronic acid, 0.2 mL on one side and 0.25 mL on the other, with aspiration showing no blood. Around fifteen hours later she developed ecchymosis, swelling and tenderness in the territory of the supratrochlear artery above the right eyebrow. She was treated with hyaluronidase alongside other measures, and a superficial epidermal erosion healed within a week.

What makes that report worth reading is not the scare value. It is that three layers of protection were in place at once: a preformed tunnel, a thick blunt tip, and a negative aspiration. All three, and it still happened.

Which sites deserve the most caution, and when a needle is still right

In that JAMA Dermatology dataset, the sites reported occluded most often were the nasolabial folds and the lips, while the greatest mean severity was at the glabella. Those two rankings differ, and the gap is informative: where events are common is not where events are worst. The glabella, nose, tear trough and temple are dangerous because they are rich in terminal branches that communicate with ocular flow. I unpack the anatomy of those zones in the vascular map of the facial danger zones.

The flip side is that "cannulas are safer" does not retire the needle. For very superficial planes, for shaping the vermilion border, for a single small deposit, a needle still offers precision a cannula cannot match. Ruling needles out across the board usually means avoiding one risk by compromising somewhere else.

Key point: The question to ask is not which device is better, but which one suits this site, this plane, this volume.

Seeing the vessel first is the real upstream step

The device decides whether a tip is more likely to pierce a vessel or push it aside. With either device, though, if you do not know where that vessel runs, you are still guessing.

Kroumpouzos and colleagues, writing on complications of lip and perioral filler (Journal of Plastic, Reconstructive & Aesthetic Surgery 2023;84:656-669, PMID: 37002059, DOI 10.1016/j.bjps.2023.01.048), place ultrasound guidance at three points: vascular mapping for prevention, telling an intravascular embolus apart from external compression by the implant for assessment, and real-time imaging of hyaluronidase delivery for management. They also stress that practitioners should be versed in injection anatomy and in the prevention, recognition and management of complications. Their focus is the perioral area, but that three-stage use of imaging transfers readily to other regions.

This is what we mean by seeing before treating. I have handled a fair number of cases that went wrong because the work was done by feel, without imaging, and after enough of those it is hard to trust tactile judgement again. Ultrasound does not drive risk to nothing. What it does is replace "I don't think there's a vessel here" with "I can see the vessel, and it is here." Device, imaging, anatomy and injection technique are bound together, and dropping one shifts the load onto the other three.

Four questions worth asking at consultation

Patients often assume they lack the expertise to ask anything useful. In fact a few questions separate the answerable from the unanswerable rather clearly:

1. For this site, are you planning a cannula or a needle, and why?
2. How are you establishing where the vessels run in this plane?
3. If an occlusion does happen, what is your first step, and is hyaluronidase in the clinic right now?
4. How do you control volume per point and injection speed?

I particularly recommend the third. A doctor willing to walk you through the rescue protocol has usually thought about it happening.

Common questions

Can vascular occlusion still happen with a cannula?

Yes. The literature includes a case report of occlusion at the glabella despite a 22G blunt tip, a preformed tunnel, and aspiration showing no blood (DOI 10.1111/jocd.70314). The risk reduction cannulas offer is clear in large-sample data, but it is a reduction, not an elimination.

If aspiration shows no blood, does that mean it is safe?

No. The conclusion from reviewing the literature is that aspiration cannot be relied on as a safety measure (PMID: 33512439). A negative draw-back is reassuring in a way the evidence does not support. Controlling the plane, slowing down and lowering injection pressure are the more substantial measures.

Do the glabella and tear trough have to be done with a cannula?

For vessel-dense sites where consequences are serious, the literature leans towards cannulas. Whether to use one, which gauge, and which plane still depends on your anatomy and the problem being treated, which is why an in-person assessment answers this better.

I have already been injected. What should send me back urgently?

Pain out of proportion during or after injection, skin blanching, a mottled or reticulated purple pattern, blurred vision or a visual field defect, eye pain. With any of these, do not wait. Contact your injector or book a clinic assessment. Occlusion management is counted in hours.

Does ultrasound need to be used on every single pass?

Not with the screen under the needle for every pass. The common approach is to map the vessels and confirm the plane for the region beforehand, then add real-time imaging at high-risk sites. What matters is knowing where the vessel is before the tip goes in.

The cannula-versus-needle argument has a less dramatic answer than people expect. Cannulas are the steadier choice in vessel-dense territory, and the data support that. They are not insurance, and the literature already records a case where all three layers of protection were in place and it happened anyway. What actually brings risk down is the person who treats device selection, imaging, anatomy and technique as one problem.

If you are considering filler, or have doubts about something already injected, you are welcome to come in for an individual assessment through our consultation page. You can also read how we approach vascular occlusion repair and filler repair.