Why does a face lose volume as it ages?
Ageing changes the face in more than one layer, and volume loss is one of them. A 2012 review of the ageing skeleton describes bone that resorbs in a specific and predictable pattern: the midface, including the upper jaw beside the nose, parts of the rim of the eye socket, and the area in front of the jowl. Less bone leaves less support for everything that sits on it.
The fat changes too. A 2012 CT study of 12 cadaver heads found that the deep fat of the inner cheek loses volume with age and that the cheek fat compartments shift downward. A 2018 study that scanned 14 living people about 10 years apart found the superficial cheek fat stayed the same size between roughly 50 and 60. Read together, they suggest that some of the change is loss and some is descent, and the two need different tools.
Filler answers the first. Hyaluronic acid is a substance the body already makes, and it holds water. Cross-linked into a filler gel, it holds its shape where it is placed and lasts longer in the tissue. Placed near bone or in the deep fat, it can replace some of what has gone. It does nothing for descent, and that limit comes back at the end of this article.
What do the trials show, area by area?
Most of the evidence comes from randomised trials that compare treated people with people left untreated for the trial period, usually rated by evaluators who did not know who had been treated. The outcome is the share of people who improved by at least one grade on a photo scale. Here is what they found, and who paid.
- Temples, 2025 trial: 80.4% improved at 3 months, against 13.5% untreated.
- Midface, 2013 trial: 85.6% of treated people improved at 6 months.
- Chin, 2021 trial: 56.3% at 6 months, against 27.5% still waiting.
- Jawline, 2024 trial: 69.0% at 6 months, against 38.0% untreated.
- Under-eyes, 2021 trial: 83.1% at 3 months, against 15.6% untreated.
- Lips, 2021 pooled analysis with no untreated group: 91% rated fuller at 2 months.
All five trials were paid for by the company that makes the gels tested. The details and caveats for each follow.
Temples. In a 2025 trial, 171 adults with hollow temples were randomised 2 to 1 to filler or no treatment. At 3 months, 80.4% of the treated group had improved by at least one grade in both temples, against 13.5% of the untreated group, and in the trial the effect lasted more than a year. Allergan Aesthetics, the filler's maker, funded it, and 83% of participants were White.
Midface. The pivotal midface trial, published in 2013, randomised 235 adults aged 35 to 65 to filler and 47 to no treatment. At 6 months, 85.6% of the treated group had improved by at least one point, rated by two blinded evaluators, and nearly half kept the correction at 24 months. Allergan, the gel's maker, ran it and submitted it to the US FDA.
Chin. In a 2021 trial sponsored by Allergan, 192 adults with a receding chin were randomised 3 to 1 to treatment at the start or 6 months later. At 6 months, 56.3% of the treated group had improved by at least one point, against 27.5% of those still waiting, and the benefit was still evident at 12 months.
Jawline. In a 2024 trial funded by Allergan Aesthetics, 69.0% of 157 treated adults had a better-defined jawline by at least one grade at 6 months, against 38.0% of 49 untreated. That second figure is a reminder of how far a rating can move with no treatment at all.
Under-eyes. In a 2021 trial supported by Allergan Aesthetics, 83.1% of the treated group improved by at least one grade at 3 months, against 15.6% of the untreated group. It comes with caveats. Of the 135 people analysed, 80% were White and 3 were Asian. The authors note that only the under-eye was treated, to suit the US registration, and that the volumes were probably larger than everyday practice uses. Three people had swelling that began more than 30 days after treatment. And the trial treated hollows, which are only one of the things people call eye bags.
Lips. A 2021 meta-analysis from Hungarian university groups, with no company funding, pooled lip studies that included cohort studies and case series. It estimated the share of treated people whose lips were rated fuller by at least one point: 91% at 2 months, 74% at 6 months and 46% at 12 months. There was no untreated comparison group, so these figures are weaker evidence than the trials above. Across 1,496 participants, tenderness (88.8%), swelling (74.3%) and bruising (39.5%) were the common effects. Rarer ones included lumps called foreign body granulomas (0.6%) and angioedema, a deep swelling (0.3%), and the authors found a small number of serious effects. They call for better randomised trials.
Three things are worth holding on to. All five trials were paid for by one company, Allergan, and each tested one of its own gels, so they show what those gels did, not what every filler does. Company-funded trials are common in this field; that is a reason to read the numbers with care, not to dismiss them. They compare filler with nothing, so they show that filler changes the area; they cannot say whether it was the right choice for a given face. And most participants were White: 80 to 83% in the temple, chin and under-eye trials, with only 3 of 135 in the under-eye trial Asian, 58% in the midface trial, and mainly White in the jawline trial. How well the results carry over to Asian faces has been tested less than it should be. In several of the trials, participants' own satisfaction scores rose as well.
Does filler have medical uses?
Yes. Some of its approved uses are medical. The US FDA lists the restoration and correction of facial fat loss (lipoatrophy) in people with HIV among the approved uses of dermal fillers. That use was approved for two fillers that are not hyaluronic acid: poly-L-lactic acid in August 2004 and calcium hydroxylapatite in December 2006. A 2015 systematic review of fillers for HIV facial fat loss gave poly-L-lactic acid the strongest grade of recommendation it awarded, a B.
Acne scars are the second. The FDA lists the correction of contour deficiencies such as acne scars among approved uses, and notes that permanent fillers are approved only for the folds beside the nose and for cheek acne scars. In a 2014 double-blind trial of 147 people with rolling acne scars, whose first author worked for the filler's maker, 64% of those treated with a permanent filler, PMMA microspheres in collagen, met the success measure at 6 months, against 33% given saline. Being permanent, it cannot be dissolved. Hyaluronic acid has been tried for scars as well: in a small 2022 split-face trial of 15 people, whose lead author consults for the gel's maker, the filler side improved more than the saline side, though saline improved a little too.
Neither HIV approval was for hyaluronic acid. The word filler covers several materials, and they behave differently, including in whether they can be dissolved.
How long does hyaluronic acid filler last?
Often longer than people are told. The FDA's page on approved fillers says the effects of cross-linked HA last approximately 6 to 12 months. A 2026 systematic review of 24 studies and 1,410 patients found more. Of the 24 studies, 64% measured volume with 3D imaging. In those, 60 to 86.4% of the gel was still there at 12 months, and 50 to 86% at 24 months. One study reported only 6 to 9 months. The authors, who declared no conflicts, tie the differences to how the gel is made, how much the area moves, and the anatomy. Two caveats come with it. The review included only studies that followed people for at least a year, which favours longer figures. And it linked one gel technology, the one tested in all five trials above, with more late side effects.
MRI adds to the picture, with limits. A 2024 Australian review of 33 people scanned after midface filler found HA visible in every one of them, including 12 who had not been injected for more than 5 years. It was a small, selected group: 9 were scanned because of swelling, and the injection histories partly relied on memory. MRI shows that gel is present. It does not show how much remains or whether it still changes the face. Its authors suggest that refill schedules be reconsidered.
The trials point the same way: the temple effect lasted more than a year, and nearly half the midface group kept the correction at 24 months. In the 2026 review, satisfaction fell as the measured volume fell, so the gel and the result fade together.
Why does overfilling happen?
If gel stays longer than expected, top-ups on a fixed schedule can stack. That is one route to the overfilled look, the "pillow face" people now fear. A 2024 review of facial overfilled syndrome names the usual causes: filler in the wrong place, the wrong product, injectors who are overzealous, and clients who "chase the lines". It adds that the syndrome is commonly produced when filler is relied on as a single tool.
A second 2024 paper describes the forms it takes, from heavy foreheads to "pillow" faces, and concludes that it predominantly affects people of East Asian descent because of their facial characteristics. Its prevention advice is to use less filler, to address the underlying bone changes, and to remember that recent imaging challenges the idea that filler simply breaks down.
So much of the overfilled look comes down to dose and placement, and the same papers describe how to lower the chance of it. If a face has had filler before, the first step is to find out what is still there, before anything is added.
Can filler be dissolved, and what are the limits?
Hyaluronic acid can be broken down by an enzyme, hyaluronidase. The enzyme works on HA only. The FDA warns that removing permanent fillers, or fillers not made of HA, may be difficult or impossible.
Even for HA, dissolving is not a clean undo. A 2024 review from Moorfields Eye Hospital in London looked at 157 eye areas in 90 people treated with the enzyme. Of those, 59% had a satisfactory result, 24% needed more enzyme, and 18% complained of changes such as hollowing afterwards. Those changes were linked to how long the filler had been in and how much there was. The enzyme concentration made no significant difference. The authors advise that people with longer filler histories and larger volumes be told the risk is higher.
Which late lumps suit the enzyme, and which do not, is set out in filler gone wrong: lumps, late swelling and dissolving. Its limit for the eye is in the next section.
How risky is filler, in proportion?
The common effects are the mild ones: tenderness, swelling and bruising, which in the under-eye trial mostly settled within a week. Serious problems are uncommon. The most useful counts come from large practices that counted their treatments as well as their problems, and most of those counts were made by looking back through the practices' own records, which tends to give the lowest figures. The trials, which followed every participant, recorded more.
- A 2011 prospective study across 8 US dermatology practices recorded 4,430 filler procedures. Adverse events followed 0.52% of them, mostly lumps. Among them were 2 blocked vessels and 2 patches of skin necrosis, where skin tissue dies, and the authors classed none of the events in the study as serious.
- One US practice of 3 board-certified dermatologists looked back over 11 years and 7,659 patients. It found 4 serious events, 3 delayed lumps and 1 blocked vessel that led to skin necrosis, and all 4 resolved without significant long-term effects.
- A 2021 study of 370 US board-certified dermatologists, reporting 1.7 million syringes from their own records, recorded about 1 blocked vessel per 6,410 syringes injected with a needle and 1 per 40,882 with a blunt cannula. Of those blockages, 85% left no lasting effect. The rest were graded as leaving a scar or an eye injury.
- A Japanese cosmetic clinic group, looking back through its own records, found 12 serious complications, 2 infections and 10 blocked vessels, in 290,307 HA injections between 2020 and 2024. Of the 10 blockages, 8 followed forehead injections.
- In the trials above, where every participant was followed up, the jawline trial recorded serious treatment-related problems in 3 of 198 treated people (1.5%), an infection, a lump and an allergic-type reaction, with 2 more lumps after a repeat treatment. In the chin trial, 1 of 192 people had 2 serious problems, an infection and a lasting inflammation, both of which resolved.
Each figure belongs to its own setting: large practices, mostly experienced dermatologists, counting from their own records, or a trial with its own rules. What counts as serious also differs: the 2011 study did not class its 2 patches of necrosis as serious, while the single-practice review counted necrosis as serious. None of them is a rate that transfers to another clinic, and none can count what was never reported. What they do show is proportion. Across these series, serious harm is uncommon. It is not impossible: in a 2016 survey answered by 52 of 127 experienced filler trainers, 62% had had at least one injection enter a blood vessel.
Vision loss is the complication most people have heard of, and it has no rate. A 2024 review counted 365 new published cases between 2018 and 2023, on top of 146 already reviewed, which is a count of reports with no denominator. Of the 318 with a recorded outcome, 68.2% had no recovery of vision, and no treatment, the dissolving enzyme included, was significantly linked with recovery. The enzyme is a tool for skin and for lumps. For sight it is no safety net. Which sites appear most often in those cases, and why, is in dermal filler risks in Malaysia.
After any filler, skin that turns white, grey or blue, pain out of proportion to the treatment, any change in vision, or any sign of a stroke, such as sudden trouble speaking or weakness of the face or a limb, needs care straight away. A change in vision or a sign of a stroke means the nearest emergency department, at once.
What does Malaysian law say about filler?
Several of the published Malaysian cases involve filler in the breast or buttock. The Ministry of Health's Medical Practice Division lists breast filler and buttock filler as illegal procedures, alongside IV drips of vitamin C. Malaysian doctors have written up what followed. In 2025 a respiratory team described 4 patients with lung complications after HA filler in the breast or buttock, one of whom died, and noted that most such procedures happen in settings where no one expects a catastrophe. In 2023 an eye team described a 27-year-old woman who lost vision after HA breast filler given by a non-medical practitioner; with treatment, her vision partly recovered, to 6/12, over 8 weeks. The law, the risks and what to do if it has already been done are in breast and buttock filler are illegal in Malaysia. The Retreat Clinic does not offer either.
Facial filler sits in a different place in law. The Ministry's Guidelines on Aesthetic Medical Practice list filler injection, excluding silicone and fat, as a minimally invasive aesthetic procedure for a registered medical practitioner who holds a Letter of Credentialing and Privileging (LCP) for it, working in a licensed or registered facility. The gel is a medical device, so it should be on the Medical Device Authority's register. Silicone is not approved by the US FDA for any aesthetic use. The three checks, the doctor, the LCP and the product, take minutes, and the steps are in how to check an aesthetic doctor in Malaysia.
What can filler not do?
Filler adds volume. It cannot tighten loose skin, and it cannot lift skin that has dropped, which is the descent the imaging studies above describe. It does not change skin quality, texture or pigment. The American Society of Plastic Surgeons notes that fillers are temporary and that for some people a facelift, brow lift or eye lift may be the better approach. Laxity has its own assessment, set out in sagging skin and jawline laxity, and how filler differs from a muscle relaxer and a collagen stimulator is laid out in the injectables comparison guide.
Asking filler to do those jobs is one of the routes to the overfilled face: more gel added to a problem that volume does not answer. And some faces, particularly younger ones, need nothing at all.
Filler has one job, putting back volume that has been lost. Much of what has gone wrong with its reputation began when it was asked to do another job, or to do its own job too often.
