Recently I saw a pie chart being shared around: PMMA 35%, PLLA 31%, CaHA 27%. The discussion underneath quickly turned into "which one is the most dangerous" and "my Sculptra is in second place, should I be worried?"
The chart comes from a 2025 systematic review in the Journal of Cosmetic Dermatology that collected cases of foreign body granuloma after collagen stimulator injections (Wang H, et al. 2025. PMID 41132036). The study itself is careful work. Using it to rank products by danger reads it in the wrong direction.
My clinic sees the other end every day: people who already have a lump and are looking for someone to remove it. So the question I care about is a different one. How many of these problems were already set up at the moment a doctor decided how much to inject?
Key point: The percentages in the 117-case review describe what the affected patients had been injected with. They cannot answer how many people who receive a given product will develop a problem.
What the review collected
The authors searched four databases for reports published up to August 2024 and included 40 publications: 39 case reports and one retrospective study, 117 patients in total. 109 were women, aged 29 to 78.
← Swipe to see more →
| Material | Cases (share) | Injection to detection (months) |
|---|---|---|
| PMMA (polymethyl methacrylate, non-absorbable microspheres) | 41 (35.04%) | 35.34 |
| PLLA (poly-L-lactic acid, the active ingredient in Sculptra) | 36 (30.77%) | 18.75 |
| CaHA (calcium hydroxylapatite, the active ingredient in Radiesse) | 32 (27.35%) | 7.35 |
| PCL (polycaprolactone, the active ingredient in Ellansé) | 5 (4.27%) | 13.20 |
| Dextran microspheres | 3 (2.56%) | 2.33 |
The authors link this distribution to the properties of the materials themselves, such as immunogenicity and degradability. Nodules were the most common presentation at 82.91%, followed by swelling at 9.40%. The most common location was the mouth and the area around it at 30.77%, with the lips a further 8.55%.
Why these percentages are not a risk ranking
There is a numerator but no denominator
The 35% in the PMMA row means that of the 117 people who developed a granuloma, 41 had received PMMA. How many people worldwide have had PMMA, and how many of them developed a granuloma, is not in this data.
To compare which material causes problems more often, you need to know how many people received each one. Case reports cannot provide that number.
Published cases are already a selection
The cases that get written up are usually the rare ones, the severe ones, the ones that dragged on, or the ones treated in a new way. People who had no problems never appear in the literature, and most problems quietly handled in a clinic don't either. Nobody knows how far 117 is from the real total number of granulomas.
The more a product is used, the more cases get reported
PLLA has been marketed worldwide for many years and is used very widely. If far more people receive it than other materials, a higher number of case reports does not mean each person who gets it carries a higher risk.
PMMA is a different story again. Its approval and use vary a lot between countries, and products called PMMA belong to different generations. The authors note that the second-generation product Artecoll, with a larger share of particles under 20 µm in diameter, had clearly more granuloma reports than the later Artefill. Pooling cases across generations and markets and then comparing them with other materials cannot tell you which is riskier.
PCL at 4.27% cannot be read as the safest either. As I wrote in the article on delayed Ellansé nodules, PCL makes up a considerable share of the stubborn lumps that need physical removal in my revision clinic. The low share in the literature and the high share in my clinic do not contradict each other. They are counted against different denominators.
The authors say so themselves
The limitations section of the review is explicit: the data come mainly from case reports, and the results cannot serve on their own as a basis for determining the incidence of these fillers. The abstract's conclusion says non-degradable fillers had a higher "reported incidence", and that word should be read as "more cases reported".
A study with a denominator looks different
Compare it with data that does have a denominator. A US multicentre chart review covered 1,002 people who received PLLA on the face, 4,483 treatments in total, reconstituted to a final volume of 8 to 10 mL including anaesthetic. Nodules were recorded for 4 people, about 0.4% (Palm M, et al. 2021. PMID 33400413). It is a retrospective chart review and only counts what was written down, so the true figure may be higher. But it answers the question "how many people who received it had a problem", which the 117-case review cannot.
Key point: To know the risk of a material, look for studies with the number of treated patients as the denominator. A collection of case reports tells you what problems look like when they happen.
What this data is actually useful for
It doesn't work as a ranking, but three things in it are genuinely useful.
The first is timing. From injection to detection of a granuloma ranged from one week to 15 years, around 20 months overall. For PMMA, which the body does not absorb, the reported figure was close to three years. Six months without trouble does not mean you are in the clear. If the same spot starts to harden or grow years later, the material injected back then should still be on the list.
The second is location. The most common site was around the mouth (30.77%), followed by the lips, between the brows (5.13%) and around the eyes (4.27%). A study the authors cite suggests particles migrate more easily in areas that move a lot with expression. The authors also note that the dermis is more immunologically reactive, so injecting into the dermis is more likely to lead to granulomas than injecting under the skin or at the dermal-subcutaneous junction.
The third is treatment outcome. Among the 117 patients, the most common treatment was intralesional steroid injection (25 patients), followed by surgical removal (18). The nodule shrank in 32, resolved completely in 12 and did not improve in 8, while 65 had no follow-up or no reported outcome. The authors observe that the stubborn cases that respond poorly to treatment are mostly resolved in the end by surgical removal.
What you can control more than the product choice: how much, how it is prepared, which layer
When the review discusses what causes nodules, it lists a series of factors: injecting too superficially, too high a concentration of microspheres, no massage afterwards, an excessive dose, too short an interval, incorrect preparation, and too little hydration time. These let particles clump together and form nodules. The authors note that these are mostly non-immune nodules caused by particle clumping, not quite the same as inflammatory granulomas, though both leave people with a lump they need dealt with.
Every item on that list has one thing in common. It is decided at the time of injection.
← Swipe to see more →
| Decided by the doctor at the time of injection | Outside the doctor's control, but worth asking about beforehand |
|---|---|
| How much per session, and how much over the whole course | Individual immune status, autoimmune disease |
| Dilution and how the product is reconstituted | Later infection, facial trauma, flu-like illness, herpes |
| Which layer it goes into (dermis or below) | Old filler already in the face that nobody can identify |
| Amount and spread at each point | The particle characteristics of the material itself |
| Massage afterwards and the interval to the next session |
The right-hand column does play a part. The authors describe trauma, infection, flu-like illness and herpes as associated with onset, without showing cause. So not every granuloma comes from injecting too much. But the left-hand column is what the doctor decides and should take responsibility for.
How I inject Sculptra and Radiesse myself (dilution, massage, intervals, and which areas I don't inject) is set out in an article on another Liusmed site, How to Avoid Sculptra and Radiesse Nodules and Overfilling. Here I'll keep to one principle: first judge how much is needed, then decide whether to split it into sessions. The number of vials comes last.
"So many vials per face": when treatment becomes a package
Some injection training today turns treatment into a formula: so many mL for this area, so many vials for a whole face, with large-volume injection packaged as a "complete treatment". These courses are sometimes closely tied to product sales and to the trainers' commercial partnerships.
I think this is a conflict of interest the aesthetic field should face openly. When more injected means more sold, the question of how much this particular person needs is easily pushed aside.
With collagen stimulators this matters even more. The effect grows over several months, and what you see right after injection is mostly water and gel. Inject the full formula in one go, add a short interval, top up because it doesn't look like enough yet, and a few months later two rounds of collagen stack on top of each other. That is what I describe in early signs of overfilling. Material that piles up in one spot is also one of the sources of nodules.
When I remove filler in my revision clinic, I look at the curves and proportions of the whole face. The goal is not to take every last bit out. Injecting follows the same eye: first see where this face needs something and how much, and stop when the proportions are right. When a patient asks for "a bit more", judging whether it is really needed is often harder than injecting another vial, and takes more experience.
If you are considering a collagen stimulator, these are questions you can ask the doctor beforehand:
- How much will be injected this time, and how was that amount decided?
- What dilution, and which layer?
- How should I massage afterwards? How long until the next session, and how will we decide whether another one is needed?
- Will you look at whatever I've had injected before, before injecting anything new?
The answers will show you how the amount was decided.
If you can already feel a lump, how to work out what it is
Whichever product you had, once a lump appears there are three things to establish first: which layer the material is in, how large the area is, and whether there is a capsule or active inflammation around it. That needs ultrasound. It is hard to judge by touch alone.
Early, active inflammation may be controlled with medication. For a nodule that is already encapsulated and stable, repeated steroid injections often leave the surrounding tissue atrophied and sunken while the lump stays. At that stage, ultrasound-guided physical extraction is a realistic option. The overall approach to lumps from different collagen stimulators is gathered on the collagen stimulator condition page.
To understand how granulomas form, see what a foreign body granuloma is. If you are not sure whether what you feel is a granuloma, leftover filler or scar tissue, start with filler lump, granuloma or scar tissue. If it was Sculptra, removal planning is covered in Sculptra granulomas and delayed nodules.
Frequently Asked Questions
Which collagen stimulator is least likely to cause granulomas?
There is currently no data that can answer this directly. The 117-case review only collected cases where something went wrong and has no figures for how many people received each material. Individual studies of each product also differ in follow-up length and how patients were enrolled, so they cannot be compared head to head. The material's properties matter, but dilution, injection depth, dose and intervals are the parts that can be controlled at the time of injection.
Is PMMA really the most dangerous?
PMMA is not absorbed by the body and stays in the tissue, so the window in which problems can appear is longer. In this review it had the most reported cases and the latest onset, close to three years. But more reported cases does not mean a higher rate. PMMA's approval and use differ a great deal between countries, and the products belong to different generations.
Nothing has happened for a year. Can a granuloma still appear?
Yes, it can. In this review, the time from injection to detection ranged from one week to 15 years, around 20 months overall. If the same spot starts to harden, grow or swell repeatedly a year or more later, it is still worth checking with ultrasound.
How many vials of collagen stimulator do I need?
There is no fixed number that applies to everyone. It depends on your facial proportions, tissue thickness, whether there is old filler already present, and how much collagen the previous session actually produced after a few months. Because the effect takes months to show, splitting treatment into sessions and reassessing after each interval makes over-injection less likely than injecting everything at once.
Can a granuloma be dissolved with steroid injections?
Some early, actively inflamed lesions can first be controlled with medication. Steroids were the most common treatment in this review, but stubborn cases that respond poorly were mostly resolved in the end by surgical removal. Repeated steroid injections can cause the surrounding tissue to atrophy and sink, so it is best to confirm what the lesion is with ultrasound before injecting.
If you want to know where you stand
If you have had a collagen stimulator and can now feel a lump, or one spot keeps swelling up, first see a doctor who uses ultrasound to examine fillers, and confirm which layer the material is in, how large the area is, and whether there is a capsule. If you haven't been injected yet and are comparing products, you are welcome to bring the questions above.
To discuss your situation, start with booking a consultation.
References
- Wang H, Wu D, Lo C, Liu S, Ji Q, Reem AA, Qiu H. Foreign Body Granulomas Reaction Related to Collagen Stimulatory Cosmetic Fillers: A Systematic Review. J Cosmet Dermatol. 2025;24(10):e70459. PMID 41132036 / DOI 10.1111/jocd.70459
- Palm M, Mayoral F, Rajani A, et al. Chart Review Presenting Safety of Injectable PLLA Used With Alternative Reconstitution Volume for Facial Treatments. J Drugs Dermatol. 2021;20(1):118-122. PMID 33400413 / DOI 10.36849/JDD.5631
