Last updated 16 August 2026 · Clinically reviewed · Hampstead Aesthetics, West Hampstead, London
You are standing in your bathroom with the morning light catching your face at an unfamiliar angle. The reflection startles you. Your cheekbones do not look sculpted anymore. They look sharp, almost severe. The soft fullness you remember has retreated, leaving hollows beneath the bone that cast shadows your fingers can trace.
You turn your head and notice the midface has flattened. This did not happen overnight, and now the change is unmistakable. It is not dramatic enough to alarm anyone else. It is enough that you wonder if others notice.
You have heard someone mention Sculptra for cheek volume in passing and you start to think there might be something you could do. But first, you want to understand what is happening beneath the skin.
The short answer
Hollow cheeks come from four changes at once. The fat compartments of the midface deflate and slide downward, the bony platform beneath them retrudes, the deep buccal fat descends rather than shrinking, and the collagen in the overlying skin thins. Skincare improves the surface. It cannot replace what has gone from underneath. Where the problem is genuine lost volume, the options that work put structure back — hyaluronic acid cheek filler for immediate correction, or a collagen stimulator such as Sculptra for gradual rebuilding across months. Which suits you depends on whether you would rather see the change today or grow it.
The biology: four mechanisms behind hollow cheeks
Hollow cheeks are not caused by a single factor. Four processes converge, and they do not all do what most explanations claim they do.
Malar fat pad descent
The face does not carry one continuous sheet of fat. It carries discrete compartments separated by septa, mapped in detail by Rohrich and Pessa in 2007, and they age at different rates rather than together. The malar pad sits in the upper cheek and gives the face its convexity. Its retaining structures loosen, it drops, and it deflates.
That combination matters more than either half alone. The fat does not vanish. It moves, and it moves downward, so a hollow appears above at the same time as a heaviness appears below. Patients usually describe it to me as the cheek sliding.
Buccal fat: descent, not disappearance
This is the mechanism most often explained backwards, and it was explained backwards on this page until today.
The buccal fat pad sits deeper, between the muscles of the cheek. The intuitive story is that it shrinks and leaves a hollow. A 2026 scoping review of the anatomical and surgical literature found close to the opposite: the buccal pad tends to increase in volume between childhood and adulthood, and may not begin to reduce until after the age of 60 or 70. What changes earlier is where it sits. It descends.
So in most people under sixty the buccal pad is not causing an upper-cheek hollow. It is more often contributing to lower-face heaviness at the same moment. I had this wrong here, and the distinction changes what you would sensibly do about it.
The bony platform retrudes
Bone is not a fixed backdrop. Shaw and colleagues measured the facial skeleton on CT across age groups and found the midface angles decrease — glabellar, pyriform and maxillary — while the orbital aperture widens. The pattern is selective remodelling rather than uniform loss.
The practical consequence is simple: the shelf holding the soft tissue forward moves back, so the same volume of fat now projects less. Note what was actually measured. Angles and apertures, not bone density. This page used to say bone density loss, which is a reasonable-sounding phrase and is not what those studies found.
Collagen and elastin thinning in the overlying skin
Collagen and elastin give skin its tension and its recoil. As they thin, the skin drapes rather than holds, so any loss underneath shows through more readily. This is the one mechanism of the four that home care can genuinely influence. It is also the smallest of the four.
They compound. Malar fat descends into a face where the bony platform has already retruded, where the buccal pad has dropped into the lower cheek, and where the skin has less tension left to disguise any of it. That is why treating one layer alone tends to disappoint. Hollow cheeks are among the most common concerns I assess, usually arriving alongside aged skin and fine lines and wrinkles on much the same timetable.
Why it varies: three accelerating factors
Weight-loss medication is the newest variable and the one I am asked about most. Over the past two years I have seen more patients arriving with rapid midface hollowing, and many have been taking semaglutide, tirzepatide or another GLP-1 agonist. The mechanism deserves care, because the popular version overstates what is known. There is no controlled evidence that these drugs strip facial fat specifically. What is established is that rapid weight loss from any cause takes fat from the face, and the face has proportionally little to give. My working assumption is that this is a rate-of-loss effect rather than a drug effect, and I would rather say that than repeat a mechanism nobody has demonstrated. The same pattern turns up after bariatric surgery and in endurance athletes. What happens to the skin itself is covered in skin laxity after significant weight loss.
A naturally lean facial structure brings the whole timetable forward. Low body fat and fine bone structure mean less cushioning to start with, so a small absolute loss reads as a large visible one. I have assessed patients in their thirties with midface contours I would more often expect two decades later, entirely because of their build. Nothing has gone wrong in those faces. There was simply less in reserve.
Smoking and cumulative ultraviolet exposure act on the fourth mechanism rather than the first three. Smoking narrows the small vessels supplying the dermis, and ultraviolet light degrades collagen directly. Neither removes fat. Both thin the covering over it, which makes existing hollowing more visible without having caused it. So when someone asks whether stopping will reverse the hollow: it will help the skin, and it will not refill the compartment.
That is three different problems wearing the same face.
If any of this sounds familiar, a skin assessment is the right starting point. There is no obligation and no assumption about treatment.
What you can do at home
Prescription topical retinoids
A prescription topical retinoid is the best-evidenced thing you can put on your skin for photoageing, and the honest framing is that the effect is real and modest. A 2025 systematic review pooled eight randomised controlled trials covering 1,361 patients and found tretinoin beat vehicle for both fine wrinkles (mean difference 0.412) and coarse wrinkles (mean difference 0.245). Those are effect sizes on a graded wrinkle scale, not proportions of people who responded. They describe a change you would notice in the mirror rather than one that rebuilds a face.
Retinoids are prescription-only medicines in the UK, so strength, frequency and suitability get decided at a consultation rather than chosen from an article. That is what our prescription retinoid skin programme exists to do. Give it twelve weeks. The common mistake is starting too often, then stopping when the skin flakes in week two.
Peptide serums
Peptide serums are the mildest option here and the most often oversold, including by me in the earlier version of this page. The evidence usually cited is Robinson and colleagues in 2005: 93 women, twelve weeks, twice-daily application, significant improvement in fine lines against a placebo control. The concentration used in that trial was three parts per million of palmitoyl pentapeptide. This page previously said five to ten per cent, which confuses the dilution of a trade solution with the peptide it delivers, and overstates it by orders of magnitude. Peptides layer happily with a retinoid, and they do less. That is the whole of it.
Facial massage and lymphatic drainage
Facial massage does something, and what it does is not collagen. Gentle directional strokes moving fluid toward the neck reduce puffiness and improve circulation, and five to ten minutes daily is enough. Some patients tell me their face looks fuller afterwards, and the honest explanation is fluid distribution rather than volume, which is why it fades across the day. It costs nothing. It will not lift a descended fat pad, and nothing applied from the outside does.
Protein and dietary fats
Adequate protein supports the collagen-synthesis pathway, and a fair number of the patients I see who are actively losing weight are eating too little of it. General adult guidance sits around 1.2 to 1.6 grams per kilogram daily. That comes from the nutrition literature rather than any skin-specific trial, so treat it as a sensible baseline rather than a proven cheek-volume intervention. If you are losing weight quickly it is the most useful item here, because it protects lean tissue while the fat comes off. It is foundational. It is not a treatment.
The limits of home care
Yes. Up to a point. The four methods above genuinely improve skin quality, thickness and appearance. If your hollowing is mild and driven mainly by sun damage and thin skin rather than by lost volume, a disciplined twelve-week routine can produce a difference you can see.
The ceiling they hit is structural. Topical treatments act on the epidermis and upper dermis. Malar descent, a retruded maxilla and a displaced buccal pad all sit below that.
Once volume has gone from underneath, skincare is working on the wrong layer.
We discuss cheek filler alongside Sculptra at assessment, and there are plenty of cases where filler is the better answer.
How Sculptra restores cheek volume
Sculptra is poly-L-lactic acid, and it does not work the way hyaluronic acid fillers work. Hyaluronic acid occupies space and draws water, so the correction is there the day it is placed, whereas poly-L-lactic acid occupies almost nothing once its carrier is absorbed and instead provokes a controlled foreign-body response, after which the fibroblast activity lays down new collagen over months. Goldberg and colleagues biopsied treated tissue and documented that deposition directly. That matters, because stimulates your own collagen is a claim made for a great many products with far less behind it.
The protocol is worth stating precisely, because the manufacturer’s and mine are not identical. The instructions for use describe one to four sessions, typically three, at a minimum of three-week intervals. I space sessions four to six weeks apart and most patients need two or three, because I would rather under-treat and add than over-treat and wait it out. Nothing looks different the week after a session. That is the treatment behaving as designed, and it is the part patients find hardest.
On duration, the figure quoted around the industry needs handling. The pivotal randomised trial ran 233 subjects against a collagen comparator with a 13-month primary endpoint, and the labelling reports results lasting up to 25 months after the last session in some subjects. That is a ceiling, not an average. This page used to say most patients need nothing further for twenty-four to thirty-six months, which is more than the evidence supports.
One regulatory point, because it runs opposite to most aesthetic devices. Sculptra was approved by the FDA through the premarket approval pathway rather than cleared through 510(k) — approval and clearance are different routes with different evidence requirements, and this is the more demanding of the two. In the UK it is a CE-marked medical device under the EU Medical Device Regulation. Its licensed indication is correction of nasolabial folds and other facial wrinkles.
Using it to rebuild cheek volume is off-label.
Widely practised, well described in the literature, and still off-label. I say so at assessment.
Sculptra is £500 for one vial and £900 for two, the standard initial treatment. Consultations are £50, redeemable against treatment or product on the day. How many sessions you need is decided after I have looked at your face. Current figures are on the Hampstead Aesthetics price list. How I plan a course, and who it does not suit, is set out on the Sculptra treatment in London page.
Replacing volume, or rebuilding it
Almost every question about hollow cheeks reduces to one parameter: whether the volume is placed or grown. Cost, timing and duration all follow from that single choice. The treatments sort themselves accordingly.
| Option | How volume arrives | Visible when | Typical duration | Suits |
|---|---|---|---|---|
| Hyaluronic acid cheek filler | Placed | Same day | Months to around two years | Correction now, or testing the effect first |
| Sculptra (poly-L-lactic acid) | Grown | 8 to 16 weeks | Up to 25 months in some patients | Diffuse deflation, and patience |
| Ellansé | Both | Same day, then builds | Product-dependent | Some immediate change plus stimulation |
| PDO thread lift | Neither — repositions | Immediate | Around 12 to 18 months | Descent where volume is adequate |
| Surgical lift | Repositions | After healing | Years | Marked descent with skin excess |
Filler is the right answer more often than a page about Sculptra would lead you to expect. If you want to see the change at the appointment, or you are not certain you want it at all, hyaluronic acid is reversible and Sculptra is not. That reversibility is a real advantage and I would not talk anyone out of it.
What Sculptra will and will not do
| Sculptra will | Sculptra will not |
|---|---|
| Rebuild diffuse midface volume gradually, through your own collagen | Correct anything on the day — nothing shows for around eight weeks |
| Improve skin quality and thickness across the treated area | Give precise, sculpted definition, which is a filler job |
| Hold longer than most hyaluronic acid products, in the patients it suits | Lift tissue that has descended, which needs threads or surgery |
| Treat temples and lower face in the same course, where indicated | Be dissolved if you dislike it, which hyaluronic acid can be |
The last row is the one to sit with. Hyaluronic acid can be broken down with hyaluronidase within a day. Poly-L-lactic acid cannot, and you wait for it to resolve on its own. For an uncertain first-time patient I will often suggest filler for exactly that reason, even though it is the shorter-lived of the two. Where the problem is jowls and sagging skin rather than lost volume, the honest answer is a lift rather than a volumiser. I will refer rather than treat.
Key takeaways
- Four mechanisms compound: malar fat descent and deflation, a retruding bony platform, buccal fat that descends rather than shrinking, and thinning collagen.
- Rapid weight loss from any cause, including on GLP-1 medication, takes volume from the face early. The evidence points to the rate of loss rather than the drug.
- A prescription retinoid has real but modest evidence behind it. Peptides are gentler still. Both work on skin quality, not volume.
- Sculptra rebuilds volume through your own collagen across months. It is off-label for cheeks, it cannot be dissolved, and nothing shows for around eight weeks.
- Filler is the better choice for some people, particularly anyone who wants the result on the day or wants to keep the option to reverse it.
Frequently asked questions
What causes hollow cheeks?
Four things at once. The malar fat pad descends and deflates, the bony midface retrudes so the platform under the soft tissue moves backwards, the deep buccal pad slides downward, and collagen in the overlying skin thins. Ageing is the usual driver. Rapid weight loss, a lean build, smoking and sun exposure bring it forward.
Can hollow cheeks be reversed?
Partly, and it depends what caused them. Skin quality responds to a prescription retinoid over twelve weeks or so, which softens the appearance without touching the volume, while genuine lost volume has to be replaced or rebuilt — filler does the first, Sculptra the second. Descent may need repositioning instead. Neither of them does that.
Do hollow cheeks get worse with weight loss?
Yes, and disproportionately. The face holds a small fat reserve relative to the body, so it gives volume up early. Cheeks, hollow temples and under-eye hollowing tend to go first. Losing weight slowly gives the skin more chance to adapt. It does not prevent the loss itself.
What is the difference between cheek filler and Sculptra?
Filler places volume. Sculptra grows it. Hyaluronic acid gel occupies space and binds water, so you see the result the same day and it can be dissolved if you dislike it, whereas poly-L-lactic acid triggers collagen production over the following months, so the change arrives slowly and cannot be reversed. Different tools, different patients. Neither is a better version of the other.
How long does Sculptra last in the cheeks?
The labelling reports results lasting up to 25 months after the last session in some subjects, and the pivotal trial’s primary endpoint sat at 13 months. Read 25 months as a ceiling, not an average. Plenty of people are back sooner. Best results usually appear twelve to sixteen weeks after the final session.
Is Sculptra better than filler for cheek volume?
No. They answer different questions. Sculptra suits diffuse deflation in someone willing to wait months for a gradual change, while filler suits defined correction you want to see immediately and carries the advantage of being reversible. If you are undecided, that reversibility is a good reason to start with filler.
Does it hurt, and what is the recovery like?
Moderately, and minimally, in that order. Most patients describe discomfort rather than pain, helped by topical anaesthetic. Redness, swelling or bruising usually settles within forty-eight hours. You will be asked to massage the treated areas at home for several days after each session. That part is not optional. It is how the product is distributed evenly.
Can Sculptra look unnatural?
It can, and the usual route to it is over-treatment. Because nothing is visible for weeks there is a temptation to add more before the first session has declared itself, and then the whole result arrives at once a few months later. Treating conservatively and reviewing at six to eight weeks is the guard against that. Nodules are the other recognised risk: uncommon, and reduced by correct dilution, depth and massage.
Arrange a skin assessment
If you have recognised yourself here, arrange a skin assessment at Hampstead Aesthetics. We will work out which of the four mechanisms is doing most of the work in your face, and what is realistic. That includes whether doing nothing is the sensible answer for now. The clinic is at 269 West End Lane, West Hampstead.
Related reading
- Why your face is losing volume
- Skin laxity after significant weight loss
- How long cheek fillers take to settle
- Temple filler for hollow temples
Sources
- Rohrich RJ, Pessa JE. The fat compartments of the face: anatomy and clinical implications for cosmetic surgery. Plast Reconstr Surg 2007;119(7):2219–2227. doi:10.1097/01.prs.0000265403.66886.54
- Bellity P, Pechevy L, Carloni R. How does the buccal fat pad age and how can it be managed in facelift surgery? A scoping review of the literature and our proposed algorithm. Aesthetic Plast Surg 2026;50(6):2205–2219. doi:10.1007/s00266-025-05300-1
- Shaw RB Jr, Katzel EB, Koltz PF, et al. Aging of the facial skeleton: aesthetic implications and rejuvenation strategies. Plast Reconstr Surg 2011;127(1):374–383. doi:10.1097/PRS.0b013e3181f95b2d
- Huang HY, Lee LTJ. Tretinoin for photodamaged facial skin: systematic review and meta-analysis of randomized controlled trials. Dermatol Pract Concept 2025;15(4):a5172. doi:10.5826/dpc.1504a5172
- Robinson LR, Fitzgerald NC, Doughty DG, et al. Topical palmitoyl pentapeptide provides improvement in photoaged human facial skin. Int J Cosmet Sci 2005;27(3):155–160. doi:10.1111/j.1467-2494.2005.00261.x
- Goldberg D, Guana A, Volk A, Daro-Kaftan E. Single-arm study for the characterization of human tissue response to injectable poly-L-lactic acid. Dermatol Surg 2013;39(6):915–922. doi:10.1111/dsu.12164
- Sculptra Aesthetic instructions for use and summary of safety and effectiveness data. Galderma Laboratories. FDA premarket approval P030050. Source of the session protocol, the 233-subject controlled phase, the 13-month primary endpoint and the 25-month durability figure. The peer-reviewed report of the same registration trial is Narins RS, Baumann L, Brandt FS, et al. J Am Acad Dermatol. 2010;62(3):448–462, doi:10.1016/j.jaad.2009.07.040.
Note on independence: sources 6 and 7 were funded or sponsored by the manufacturer of Sculptra, as declared in the papers. They are the primary evidence for the treatment, and you should weigh them knowing that.
Disclaimer
This article is general information and not a substitute for a medical consultation. Sculptra and dermal fillers are offered at Hampstead Aesthetics, so treat this as an explanation from a clinic that sells the treatment. Using Sculptra to rebuild cheek volume is off-label. Outcomes vary between individuals.
About the author
Dr Chiké Emeagi is the founder and medical director of Hampstead Aesthetics, with over twelve years in aesthetic medicine. MBBS, MRCGP, MRCS, MSc, MBA. Registered with the General Medical Council under number 4418489, verifiable on the GMC register. He writes and clinically reviews every treatment guide on this site, and every injectable treatment described here is delivered by a doctor rather than delegated to a technician. The clinic is at 269 West End Lane, West Hampstead.
Next steps
If you have recognised yourself in this article, arrange a skin assessment at Hampstead Aesthetics. We can discuss what is driving your hollow cheeks and what is realistic for your face.