The Face You Were Taught Is an Average
Every injector is taught a map. The facial artery runs here, the supratrochlear there, and the danger zones are the places you stay out of. The map is not wrong. It is an average, and the difficulty with an average is that it describes a population rather than the person on your couch.
This matters more than it sounds. An injector who believes they avoided the danger zone has made a claim about a specific face using a diagram drawn from other faces. The anatomical literature is unusually clear about how often that claim fails.
The commonest course is often a minority finding
A systematic review and meta-analysis of Doppler studies of the facial artery found that the most frequent course relative to the nasolabial fold accounted for under half of the arteries examined, with the second commonest pattern accounting for roughly another fifth — though that second figure pooled only a handful of studies, and its confidence interval is wide enough to span almost anything [1]. Read that the way a clinician should: the single most likely arrangement is not what most vessels do. It is what a plurality do — and the artery on the other side of the same face was counted separately.
A plurality is a poor foundation for a safety decision. It is the difference between "the artery is here" and "the artery is here more often than anywhere else, and something else is true of most people".
One caveat belongs here, because it cuts both ways. Course "types" are categories drawn by investigators, and how many categories a study draws determines whether any one of them can reach a majority. Two adjacent types may sit close enough together to fall inside the spread of a single deposit, while another type may cover a broad band. So the percentages are not a map of where the artery is. They are evidence that the classification most injectors carry in their head is coarser than the anatomy it describes.
Sometimes the textbook pattern is the rare one
Cadaveric work on the perioral region makes the point more sharply. In one series, the inferior labial artery followed the pattern most injectors would recognise as typical — running at the vermilion border of the lower lip — in a small minority of specimens, while a horizontal pattern running through the middle of the lower lip was found in roughly seven specimens in ten [2].
The teaching diagram and the commonest anatomy were, in that series, different things. An injector who has internalised only the first may be carrying a model that is not merely imprecise but inverted for that region — though this is one embalmed cadaveric series in one population, and prevalence figures from fixed tissue should not be read as settled numbers for the living face in front of you.
Depth belongs to the patient as much as to the region
The same problem applies vertically. Doppler measurement of the supratrochlear artery along the glabellar lines found that the artery sat further from both the skin and the underlying bone as body mass index rose, and further from the skin — but not from the bone — as age rose [4]. Depth is therefore not a property of the region that can be learned once and carried between patients; it is a property of the person, and it moves in a direction that a practitioner will meet constantly in an ageing aesthetic population.
The direction of that second finding surprises people, because the intuition runs the other way: tissue thins with age, so the vessel should come closer to the surface. It does not, and a reader who assumes the paper was misread will discount everything after it — so it is worth stating plainly that depth increased.
Depth is not fixed by the procedure either. A vessel measured at rest in an undistended face is not at that depth once product has been placed beneath it, the skin has been stretched, or the patient has been asked to frown. The figure in a paper is a starting condition, not a constant.
The same study is worth citing for what it did not find: no significant difference between the mean depths measured on the right and left sides [4]. That result is easy to over-read, and over-reading it would repeat the exact error this article exists to attack. A comparison of group averages cannot tell you whether any individual is symmetrical — a cohort in which half are deeper on one side and half on the other produces the same null. What it establishes is that there is no systematic side bias to build into your planning. Whether the face in front of you is symmetrical at a given point stays an open question at the chairside.
Symmetry is a question, not an assumption
Where side-to-side symmetry does break down, it breaks down by region. The perioral cadaveric series documented substantial variation between specimens, and found the superior labial arteries markedly asymmetrical between the two sides of the same cadaver [2], while the glabellar study found no detectable left–right difference in how deep its vessel sat — a failure to find a difference rather than a demonstration that none exists, and silent on whether the vessel's course is symmetrical [4]. Both are true. The practical reading is that symmetry is something to check per region and per vessel — and that a null result in a population study is not permission to assume it in a person.
The vessel under the needle is not the only vessel that matters
A review of the supratrochlear artery describes considerable variability not only in its course and branching but in its connections, including clinically significant anastomoses linking the intracranial and extracranial circulations by way of the angular, supraorbital and dorsal nasal arteries [3].
This is why distance is a weaker defence than it feels. A vessel some way from the needle can be continuous with the vessel at the needle, and the network — not the individual artery — is what carries material somewhere it should not go. It is also the anatomical reason that some of the gravest complications in this field appear in territories the injector was not treating, even though the commonest ones declare themselves locally.
The evidence has its own error bars
An article about the limits of anatomical certainty owes the reader some scepticism about its own sources, and this next part is reasoning rather than a sourced finding. The studies above are not measuring the same thing under the same conditions. Cadaveric series describe unpressurised vessels in tissue that embalming and head position have already moved, without muscle tone. Doppler series describe living, pressurised vessels — but depth measured through a probe depends partly on how hard the probe is pressed. Neither is wrong. They are different instruments reading a moving target, and not every difference between studies is a difference between patients.
The same applies to who was measured. Much of the imaging and dissection literature in this field comes from a small number of centres and a narrow range of populations. Carrying a distribution derived from one population onto a different face is the same error as carrying a diagram onto a person, one step further out and easier to miss.
The face in front of you may not be in any of these studies
Almost every distribution described here was measured in a face that had not been treated. The patient booked for a revision has had product placed repeatedly over several years, possibly threads, possibly surgery. Prior filler displaces tissue and can displace vessels with it; fibrosis changes the planes a cannula will follow and changes how a deposit spreads once it leaves the tip; scar changes where resistance is felt, and therefore what your hand is telling you.
This is the sharpest version of the article's own argument. If an average is a poor description of an untreated face, it is a worse description of an altered one — and the altered face is a growing share of aesthetic practice. Treatment history is anatomical information, and it belongs in the consultation rather than in the aftermath.
What landmarks are actually for
None of this makes surface landmarks useless, and the studies that establish them do not claim they are. Cadaveric work intended to define landmarks for locating the facial artery describes a course that deviates at least twice along its path [5]. That is what a landmark can offer: it narrows where a structure probably is, between deviations whose positions themselves vary.
So use landmarks as what they are — a way of reducing uncertainty, not of removing it. The clinical habits that follow from treating anatomy as a distribution are the ones widely recommended, and they make more sense once the reason is clear: move the instrument rather than depositing while stationary, deliver slowly and under low pressure, use small amounts, and stop at the first sign that something is wrong rather than at the point where you are certain.
Where imaging fits, and where it does not
Ultrasound can show the artery in the patient in front of you, and the meta-analysis above exists because it does [1]. It is worth being precise about what that establishes: these studies demonstrate that the vessel can be visualised and that its position varies. They are not outcome studies, and they do not show that scanning before injecting reduces the rate of vascular complications.
That distinction is worth holding onto in a market that is beginning to sell ultrasound as a safety guarantee. Imaging replaces an assumption with an observation, which is a real gain in what you know — though not, on this evidence, a demonstrated gain in what happens to the patient. A probe in untrained hands replaces an assumption with a different assumption, held more confidently.
What this changes on Monday
Less about where the needle goes than about what you are entitled to believe while it is in. Four things follow from treating anatomy as a distribution, and they are specific rather than general.
Your prior is worse than you think. If the commonest course is a minority finding, then before you touch the face the odds that this vessel sits where the diagram puts it are worse than even. That argues for choosing the plane with the fewest vessels in it, whatever the course, rather than for steering around a line drawn from memory.
The lower lip may not be safest in the middle. If a horizontal mid-lip course is the commoner pattern in the perioral series above, then the part of the lower lip many injectors treat as its safe body is the part more likely to contain the artery. That is a change of habit, not of attitude — and one to confirm against the anatomy taught on your own course rather than to adopt from an article.
"I stayed superficial" is a weaker claim in some faces than others. The distance from skin to the supratrochlear artery rose with age and with body mass index. In a thin, young glabella, staying superficial buys less clearance than the same technique buys in an older or heavier patient. The sentence you tell yourself should not be the same sentence in both.
Your confirmatory tests are probabilistic too. A negative aspirate, an absence of blanching, a familiar tactile feel — all are evidence, and all are weak. If you have understood why the map is a distribution, you have understood why these are as well. They can raise your confidence; none of them licenses a larger or faster deposit.
There is also something to change in what you say out loud. "I avoid the danger zones" is a claim this literature does not support, and it is a sentence patients hear often. The accurate version — that vascular events are uncommon, that technique reduces but does not eliminate them, and that you are equipped and prepared to manage one — is both truer and a considerably better position to be standing in if it ever happens.
Which leaves the honest answer to the question trainees ask constantly: how much anatomy do they need before they are safe. The answer is not a quantity of anatomy. It is a working relationship with uncertainty, and knowing what happens in the ten minutes after you stop — which is learned beside someone who has managed the thing going wrong, not from a diagram or from an article about diagrams.