Radiesse is calcium hydroxylapatite, or CaHA — microspheres of a calcium-and-phosphate mineral, 25 to 45 microns across, suspended in a gel carrier. The gel provides volume on the day and is absorbed over the following months; the microspheres stay behind and prompt the body to build its own collagen around them.
Fillers like this are Class III devices, approved through the FDA's premarket approval route. That route required clinical evidence for a specific cosmetic indication in a specific area, which is why filler approvals are written area by area.
Radiesse's US record is:
22 December 2006, PMA P050052: "Subdermal implantation for the correction of moderate to severe facial wrinkles and folds, such as nasolabial folds."
22 December 2006, PMA P050037: Restoration or correction of the signs of facial fat loss in people with HIV.
4 June 2015, supplement S049: "hand augmentation to correct volume loss in the dorsum of the hands."
1 September 2021, supplement S129: Jawline contour, for Radiesse (+) Lidocaine.
31 March 2026, supplement S162: Correction of décolleté wrinkles, at a 1:2 dilution with saline.
Marionette lines are not named in any of them: Nor is any other named landmark of the lower face other than the jawline.
This is the part most pages get wrong in one direction or the other.
The 2006 indication is not written as a list of permitted lines. It says "moderate to severe facial wrinkles and folds, such as nasolabial folds." That parenthetical is an example, not a boundary. Marionette lines — the creases running down from the corners of the mouth toward the chin — are facial folds, and they plausibly sit inside the literal sentence.
So the honest answer is not "this is off-label" and not "this is approved". It is this: whether a given marionette treatment falls inside the approved wording is a matter of interpreting the label, and the FDA approval documents do not resolve it. What is not open to interpretation is the evidence behind the wording.
Every trial the FDA reviewed for that indication was run in nasolabial folds: The regulator saw nasolabial fold data and approved wording broad enough to reach beyond it. That is normal, and it is also the reason a reader asking specifically about marionette lines will not find a trial answering their question.
Off-label use of approved products is legal and routine across medicine. Flagging it is not an accusation. It simply tells you which claims have trial data behind them and which do not.
The pivotal trial: The study behind the 2006 approval was split-face: 117 subjects, each with Radiesse in one nasolabial fold and a collagen filler in the other, ages 31 to 76. It met its non-inferiority endpoint on the Lemperle Rating Scale at 3 months. At 6 months, 78.6% of the Radiesse-treated folds scored at least 1 point better than the control fold in the same face. The population was 89.7% female and 87.2% white.
The long-term extension: Bass and colleagues followed 102 of those 117 subjects for up to 39 months after their last Radiesse injection. 40% of folds assessed at 30 months or more were graded "improved" or better. The authors themselves noted the improvement was "substantially less at this later time point." Read it carefully: 40% still improved at 30 months is not the same as "lasts 30 months," and 60% were not. No nodules, granulomas or infections were reported in those 102 people over three years.
The only published work that measured marionette lines directly used hyperdilute Radiesse across the mid and lower face and found reduced marionette-line depth on 3D imaging at day 150 (p=0.0019) in 22 charts - a retrospective review with no control group, graded Level of Evidence 4 by its own authors, from an author group with Merz ties, at a dilution that is not part of any approved labelling. It is a signal, not a result.
This is the fact that changes the calculation in the lower face.
Hyaluronic acid fillers — Juvederm, Restylane, Belotero, Versa, RHA — can be broken down with hyaluronidase, an enzyme that cleaves hyaluronic acid. Hyaluronidase has no target in calcium hydroxylapatite. It does nothing to the mineral microspheres and nothing to the carboxymethylcellulose gel they sit in. There is no approved reversal agent for Radiesse, and the current label mentions none. A 2026 review of CaHA named the absence of a reversal agent as a known limitation of the material.
Sodium thiosulfate is sometimes discussed online as an antidote. It is not an established reversal agent for Radiesse - the published evidence is contested and is set out in full on the page about whether Radiesse is permanent.
What is realistically available if the result is wrong: waiting for the material to degrade, saline infiltration with mechanical disruption through a cannula for nodules, intralesional steroids for inflammatory lumps, and rarely excision. The label states that safety beyond 3 years in the face has not been investigated — that is a study window, not a duration claim.
Every filler carries a risk of vascular occlusion: Material entering or compressing an artery, cutting off blood supply. It is rare and it is a medical emergency. The nasolabial and perioral territory is supplied by the angular artery, which connects onward to the circulation that feeds the eye, so both skin death and vision loss have been reported after injection in this region.
A review of a century of published filler blindness cases found 511 cases, of which 365 were reported between 2018 and 2023. In those 365, the material was hyaluronic acid in 79.6% of cases and calcium hydroxylapatite in 1.6%. Of 318 cases with a recorded outcome, 6.0% recovered completely, 25.8% improved partly, and 68.2% did not recover. Those are counts of published cases with no denominator, so they cannot be turned into a personal risk figure — but they establish that the outcome is usually permanent.
Two further points that get skipped. In a survey of 52 experienced injector-trainers, 62% reported having had at least one intravascular injection, and pain was mild or absent in 47% of those events. Pain is not a reliable alarm.
Lumps are the other area-specific issue. In the nasolabial pivotal trial the nodule rate was 0.9%, one patient, the same as the control side. But a consensus review of CaHA advises avoiding highly mobile areas and thin-skinned, unforgiving areas because nodules are more common there, and the Radiesse label states plainly that safety and effectiveness in the lips have not been established, with published reports of nodules there. The marionette region sits close to that mobile perioral territory.
Questions worth asking at a consultation:
Are my marionette lines mostly a fold, or mostly loss of volume higher in the face?
Why Radiesse here rather than a hyaluronic acid filler that could be dissolved?
What dilution are you using, and is that dilution part of the approved labelling?
What is the plan if I dislike the result at three months?
Do you stock the drugs and have the referral pathway for a vascular emergency?
Go to an emergency department or call 911 for any of these, during or after the appointment. These can indicate that filler has entered a blood vessel, and time matters in hours.
Any change in vision — blurring, a dark patch, double vision or loss of sight in either eye
Sudden severe pain, or pain out of proportion to the procedure
Skin turning white or pale (blanching), then dusky, purple or mottled
Weakness, drooping, slurred speech or any other stroke-like symptom
Swelling keeps growing after the first few days instead of settling
A lump appears and does not settle, or a lump appears weeks or months later
Redness spreads outward from the injection site, or the area becomes warm and tender
Tenderness, firmness or swelling appears weeks after treatment, which can indicate a delayed inflammatory reaction or infection
The correction looks like it has shifted position