1. Introduction — A Question of Depth

Every injector eventually meets the same question, and most of them meet it the wrong way round. They ask, "Where is the artery?" The better question — the one that actually keeps a patient's retina perfused on a Tuesday afternoon in clinic — is, "Which depth am I in, and what is the biggest vessel that could live there?" That is the whole game. Names of arteries are a luxury. Layers are a necessity.

Think of the face as a four-storey building. The dermis is the attic — tight, well-lit, mostly safe, the vessels up there are capillaries the width of a red blood cell on its side. The subcutaneous tissue is the second floor, where the plumbing gets serious and the named branches of the facial artery actually run. The SMAS and the deep fat pads are the basement utilities. The periosteum is the foundation. The needle you are holding is the lift. The only honest question, before you press the plunger, is which floor the doors just opened on.

We are not surgeons. We do not open the skin and trace the angular artery as it deviates around someone's left nostril for the third anatomical variant in a row. We are not interventional radiologists; most of us are not running a high-frequency probe over a patient's nasolabial fold before each pass. We work blind, on a moving target, with anatomy that varies between humans and between sides of the same face. So we trade anatomical certainty for something more useful: a probabilistic, layered map of what can possibly be in the vessel we cannot see. That map is depth.

This is a review of injection anatomy from that vantage point. It is written for the practitioner who has to make a decision, in real time, about which tool goes in which plane, and what the worst-case vessel calibre is if they are wrong about either.

2. The Layers and What Lives in Them

House would have called it differential diagnosis for a needle tip. Where is the tip? What's the worst thing it could be in? Treat the worst thing first.

Layer 1 — Papillary and reticular dermis. The vessels here are the terminal vascular tree: capillaries in the papillary dermis with luminal diameters in the order of 5–10 microns, and sub-papillary plexus vessels rising into the low tens of microns (Braverman, 1989). Reticular dermal arterioles are larger but typically remain well below 100 microns. A 30G needle in true dermis is, mechanistically, in a vascular bed that physically cannot accept and propagate a filler bolus into the retinal circulation. This is the plane in which intradermal skin boosters, polynucleotides and microbolus techniques operate. Risk here is not zero — blanching, biofilm, granuloma — but vascular occlusion of a named artery is a different category of event.

Layer 2 — Subcutaneous fat. Once the tip drops out of the dermis it enters a different vascular world. The branches of the facial artery, the angular artery, the lateral nasal artery, the superficial temporal artery and its frontal branch all run, predominantly, in the subcutaneous plane (Cotofana & Lachman, 2019; Scheuer et al., 2017). Reported luminal diameters for the facial artery and its named branches generally fall in the 1.0–2.5 mm range in cadaveric series, with the angular artery measuring approximately 1.0–1.5 mm at the level of the nasolabial fold (Yang et al., 2014). A 27G or 25G needle, advanced confidently into subcutis, is fully capable of cannulating any of these.

Layer 3 — Sub-SMAS and deep fat compartments. The deep fat pads (deep medial cheek fat, sub-orbicularis oculi fat, Ristow's space pre-maxillary) sit on bone or between SMAS and periosteum and are, in many mid-face zones, vascularly quieter than the subcutis above them. This is the anatomical rationale for supra-periosteal bolus placement in cheek augmentation: the major named arteries are above you, not where the bolus is going (Surek et al., 2015).

Layer 4 — Supra-periosteal / on bone. Pressed against periosteum, the needle is below the principal arterial layer in most mid-face zones. This is the safest single-plane bolus technique that exists in the non-surgical injector's toolkit, with two important exceptions: the glabella, where the supratrochlear artery emerges from the bone, and the temple, where the middle temporal vein and deep temporal arteries make the supra-periosteal plane its own danger zone rather than its own refuge.

3. Tool in Hand, Layer Underneath

If you adopt the layered model, tool selection becomes a deterministic consequence of plane selection rather than a stylistic preference.

A 30G or 32G short needle, used intradermally, stays in the layer where vessels are microns wide. The lift only opens on Floor 1. That is the appropriate tool for skin boosters and superficial microbolus work.

A 27G or 25G sharp needle, used for supra-periosteal bolus on solid bone, deliberately skips the dangerous middle floor and lands on the foundation. The risk window is the transit, not the destination. Slow, perpendicular insertion with continuous tactile feedback for bone, then aspirate, then deliver — that sequence exists because the only floor you are afraid of is the one you pass through on the way down.

A blunt-tip cannula, typically 25G or 22G, is the tool of choice for any volumising work that has to be performed in the subcutaneous plane itself — the nasolabial fold, the jawline, the tear trough's lateral extent. Cannulas reduce, although do not eliminate, the probability of arterial penetration, principally because the blunt tip preferentially deflects off vessel walls rather than transecting them (van Loghem et al., 2018). The cannula is not magic. It is a probability shift. If the cannula tip enters a vessel, the consequences are identical to a needle.

The general rule the layered model produces is this: superficial work, sharp short needle; deep-on-bone work, sharp needle with perpendicular entry and bone contact; everything in the middle floor, cannula by default. "I always use needles" and "I always use cannulas" are both wrong, for the same reason — they refuse to look at which floor the lift has stopped on.

Video evidence · HSI clinical faculty

4. The Five High-Risk Zones, Through the Depth Lens

Glabella. The supratrochlear and supraorbital arteries emerge from their respective foramina and run superficially within a few millimetres of skin. There is, in effect, no safe deep plane here, because the artery is already shallow. The published rate of vision loss attributable to glabellar filler is disproportionate to the volume injected in this zone (Beleznay et al., 2019). Layered reading: this is a region where the danger floor is also the obvious floor.

Nasolabial fold. The angular artery courses through the subcutaneous plane and shows substantial anatomical variation in its medial-to-lateral position. The pre-periosteal plane is comparatively safe; the mid-subcutaneous plane is comparatively not. Cannula, deep, retrograde linear threading is the technique most consistent with the layered model (Lee et al., 2014).

Nasal dorsum. The dorsal nasal artery anastomoses with the ophthalmic artery, providing the most direct retrograde route to the central retinal artery of any zone on the face. The dorsum has almost no fat, the artery sits superficially, and the local circulation is end-arterial. There is no genuinely safe depth here for a beginner; this is the zone in which the layered model says "refer or refuse".

Temple. The supra-periosteal plane is not safe in the temple because the deep temporal arteries and the middle temporal vein live there. The pre-periosteal plane is also crowded with the superficial temporal artery and its frontal branch. Safer technique here is a single perpendicular bolus on bone in the lateral temporal recess where the named vessels are statistically less likely to run (Sykes et al., 2018), with active aspiration and slow, low-pressure delivery.

Lips and perioral region. The labial arteries run, on average, sub-mucosally rather than intramuscularly, though variation is wide. Submucosal sharp-needle technique places the tip in the same plane as the artery; cannula in the deep submucosal plane shifts the probability. Vascular events here are less likely to blind a patient but routinely produce necrosis of the vermilion and philtral columns.

5. What We Cannot Have, and What We Therefore Need

Two adjacent specialties have anatomical certainties we do not. The surgeon can incise, dissect, and visually identify the second or third arterial variant in a given patient before transecting it. The interventional radiologist or the aesthetic ultrasonographer can place a high-frequency probe over the zone and watch the needle approach the vessel in real time. The aesthetic injector working without ultrasound has neither. We cannot know which variation of the facial artery any given patient has on the day they sit in our chair.

What we can know — and what is therefore the practical substitute for surgical exposure — is the size of the vessels that live in each plane, the probability distribution of where named arteries run within those planes, and the matching of tool to plane that minimises the chance of crossing one. That is what "injection anatomy" actually means in non-surgical practice. It is not the memorisation of every named branch. It is the disciplined translation of depth into risk.

The reasonable injector therefore makes three commitments before every treatment: name the plane out loud (or in your head) before the needle moves, choose the tool that belongs in that plane, and rehearse, in advance, what you will do in the first sixty seconds if the floor you are on turns out to contain an artery you did not expect. Hyaluronidase drawn up before the bolus is delivered. A written vascular protocol on the wall. The patient warned, in advance, that an emergency reversal protocol exists and may be used.

6. Conclusion

Injection anatomy, stripped of jargon, is a question about which layer the needle tip is in and what the largest vessel in that layer can be. Capillaries in the dermis are five to ten microns wide and cannot do meaningful harm. Named arteries in the subcutaneous plane are one to two and a half millimetres wide and absolutely can. Tool selection is the disciplined consequence of layer selection — short needle in the dermis, cannula in the subcutis, perpendicular needle on bone — with the well-known exceptions in the glabella, the nasal dorsum and the temple, where the rules bend because the anatomy does.

We are not surgeons and we are not radiologists. We do not get to see the artery before we cross it. What we do get to do is choose the floor we stop on, match the tool to the floor, and rehearse, in cold blood, what happens if we are wrong. Done that way, dermal filler practice is not riskless — but the risk becomes a function of decisions we control rather than anatomy we cannot see.