What Cellulite Actually Is: The Anatomy Nobody Explains

Cellulite has an address. It is not distributed through the fat, and it is not a substance that accumulates. It lives in a specific stack of tissue: the dermis, the two fat layers beneath it, and the two sheets of fascia that separate them, all laced together by a scaffold of collagen strands called fibrous septae. Almost everything worth knowing about cellulite is a statement about that scaffold.

The scaffold is why almost nothing works. It is collagen, it sits several millimetres down, and it cannot be rubbed, drunk, or sweated away — which is the fastest test for whether a claim about cellulite is plausible before you look at any evidence for it.

What is cellulite, anatomically?

Cellulite is a change in the topography of the skin caused by an imbalance of forces at the junction between the dermis and the fat beneath it. Fat lobules push outward. The dermis pushes inward. Collagen septae tether the dermis downward at fixed points. Where a strong septum holds while the fat around it bulges, you see a dimple.

That is the architectural model, and it comes from anatomy rather than from marketing. In the gluteal region, cadaver dissection has described five distinct layers — dermis, superficial fat, superficial fascia, deep fat, deep fascia — crossed by two kinds of septae. Short, thin septae connect the superficial fascia to the dermis; they are numerous but individually weak. Long, thick septae run from the deep fascia all the way to the dermis; they are fewer but much more stable. The dimple forms at the inflexible thick attachment while the softer surroundings protrude (Rudolph et al., Plastic and Reconstructive Surgery, 2019; summarised in Gabriel et al., Aesthetic Surgery Journal Open Forum, 2023).

Imaging supports this. In a 2009 MRI study, Hexsel and colleagues scanned 30 women with cellulite on the buttocks and compared an area with a depression against a smooth area on the opposite buttock in the same person. Fibrous septa were visible under 96.7% of the depressions. Every septum found ran perpendicular to the skin surface. Mean septal thickness was 2.18 ± 0.89 mm in dimpled areas versus 0.27 ± 0.64 mm in smooth areas. One reporting detail worth stating: the published abstract gives those thickness values as bare numbers with no unit attached; the millimetres come from Bass and Kaminer’s tabulation of the same data.

Note what that study did not find. Summarising it, Bass and Kaminer report that there was no relationship between septal thickness and cellulite severity. The septae are where dimples happen; septal thickness alone does not predict how bad it looks.

This mechanism has been described independently by dissection, by histology and by MRI — three methods that do not share a failure mode.

Why do women get cellulite and men almost never?

Women get cellulite and men almost never do because of how the connective tissue is built, not because of body fat. Female septae tend to run perpendicular to the skin, forming tall columnar fat chambers that can protrude upward. Male septae crisscross at roughly 45 degrees, forming a net rather than a set of pillars — and men have more of those connections, each of them physically stronger.

The strength figure is the most concrete evidence in the field, though its spread is wide: the male mean of 38.46 N carries a standard deviation of ±26.3 N across just ten donors. Rudolph and colleagues took full-thickness gluteal slices from 10 male and 10 female body donors (mean age 76, mean BMI 25.3) and pulled until the septa between dermis and superficial fascia failed. Mean breaking force was 38.46 ± 26.3 newtons in men and 23.26 ± 10.2 newtons in women (p = 0.021). The same study counted subdermal fat lobules: 10.05 ± 2.3 in men versus 7.51 ± 2.7 in women (p = 0.003), meaning men have more, smaller compartments and therefore more anchor points per unit of skin.

Dermal thickness contributes as a containment layer. Ultrasound measurement in 140 adults found men’s thigh dermis averaged 1.89 mm against women’s 1.65 mm (p = 0.003), while the subcutis underneath was thicker in women (21.3 mm versus 17.9 mm). That cohort was made up of people with diabetes, studied for injection-site reasons, so treat the absolute numbers as indicative rather than as a population norm — but the direction is consistent with the anatomical literature.

Sources: Rudolph et al. 2019 (septal force, lobule counts); Nürnberger & Müller 1978 and subsequent pathology series (orientation); Derraik et al. 2014 (dermal thickness).

One important correction to the tidy version of this story. A 2002 MRI and spectroscopy study by Querleux and colleagues reconstructed the septal network in three dimensions and concluded that it “couldn’t be modelled simply as perpendicular planes for women and tilted planes at 45 degrees for men.” The real network is tortuous. Women with cellulite had a higher percentage of perpendicular septae, not exclusively perpendicular ones. The 45-degree diagram you see in clinic brochures is a simplification of a messier reality.

Are trapped toxins a real cause of cellulite?

No. No toxin has ever been identified or measured in cellulite tissue anywhere in the peer-reviewed literature, and the one study that directly measured fat metabolism and blood flow inside dimpled skin found nothing abnormal there. Detox is a marketing frame, not a described physiological process.

The most direct test came in 1998, when Rosenbaum and colleagues at Rockefeller University studied seven adults using microdialysis probes to measure subcutaneous fat metabolism and regional blood flow in vivo, plus full-thickness wedge biopsies. They compared dimpled and smooth sites within the same thigh of the same person. Their finding: “No significant differences were noted in subcutaneous adipose tissue morphology, lipolytic responsiveness, or regional blood flow between affected and unaffected sites.” What they did find was the connective tissue difference — a discontinuous, irregular subdermal connective layer in women, smooth and continuous in men.

Seven subjects is a very small study, and only four of them had cellulite — so the within-person dimpled-versus-smooth comparison rests on four people. It should be described that way. Its design is nonetheless unusually clean, because each affected person acts as their own control.

Querleux’s MRI spectroscopy work adds a second negative. It measured water fraction inside the fat lobules and found no difference related to sex or to the presence of cellulite, explicitly stating that it “did not confirm the hypothesis of increased water content in the adipose tissue of women with cellulite.”

Where sources disagree: a vascular-and-inflammation hypothesis remains alive in the review literature. The Gabriel 2023 review sets it out in full — altered precapillary arteriolar sphincters, increased capillary permeability, hyperpolymerised glycosaminoglycans pressing on capillary walls, fluid leaking into the interstitial spaces between fat lobules, and the resulting oedema and hypoxia driving neovascularisation and sclerosis of the septae. Avram’s 2004 review lists vascular changes and inflammatory factors among the four leading hypotheses. Bass and Kaminer do not discuss this hypothesis at all. So the honest summary is: fluid dynamics may play a role in how septae become fibrotic over years, but no one has measured a toxin, and the fat in a dimple is metabolically ordinary.

How is cellulite actually graded?

Cellulite is graded on one of two scales. The Nürnberger–Müller scale (1978) sorts skin into four grades, 0 to III, using a pinch test and a standing assessment; it is simple and universally cited but was never formally validated. The Hexsel Cellulite Severity Scale (2009) scores five separate features from 0 to 3 for a total of 0 to 15, and was validated statistically.

The Hexsel Cellulite Severity Scale scores: number of depressions (0 = none, 3 = ten or more), depth of depressions, morphology of surface change (1 = orange peel, 2 = cottage cheese, 3 = mattress), skin laxity, and the Nürnberger–Müller grade itself. Totals of 1–5 are classed mild, 6–10 moderate, 11–15 severe. Its published validation reported intraclass correlation coefficients of 0.881–0.922 and Cronbach’s alpha of 0.851–0.989 across raters.

The CR-PCSS and PR-PCSS rows carry the practical test. Those two scales were developed by a team in which six of thirteen authors were employees of Endo Pharmaceuticals, the company marketing the injectable cellulite treatment the scales were built to measure. Their validation work established that a one-point change is the smallest difference that means anything clinically — so any provider quoting an improvement should be able to say which scale, rated by whom, and by how many points. That threshold is useful; the authorship is a reason to read any result scored on it with the sponsor in mind.

There is also a measurement problem worth knowing. A 2015 study of 26 women compared dermatologist photo scoring against ultrasound and cutometry and found poor correlation between the instrument readings and the clinical scores. Machines and eyes do not currently agree about cellulite severity.

How common is cellulite, and is it a disease?

Reviews consistently estimate that 80–90% of post-pubertal women have cellulite. Estimates for men run from about 2%, the figure in the 2023 Gabriel review, to under 10% on Cleveland Clinic’s patient page. Neither rests on a population survey. Cellulite is not a disease. It carries no health risk and appears in the medical literature mainly as a cosmetic and psychosocial concern.

The prevalence range should be read with its caveat attached. Some reviews quote 85–98% instead. The 2023 review by Gabriel and colleagues opens its epidemiology section by saying that “very little is known about the prevalence and incidence of cellulite” and that the 80–90% figure, while widely reported, rests on a paucity of robust epidemiological data. Nobody has run a properly sampled population survey. The number is a well-repeated estimate, not a census.

On the disease question, the founding paper is unambiguous. Nürnberger and Müller’s 1978 study of biopsies from 150 cadavers and 30 living women is titled “So-called cellulite: an invented disease,” and its abstract describes “the essential normality and inevitability of it in women” and “the near futility of treating the non-disease.” Nearly fifty years later, Cleveland Clinic’s patient page still states that cellulite “doesn’t affect your overall physical health, and it doesn’t hurt.”

Men who do develop cellulite usually have an androgen-deficient state — after castration, in hypogonadism or Klinefelter syndrome, or during oestrogen or antiandrogen therapy for prostate cancer. That is the context in which Gabriel’s review places its roughly 2% figure, and the pattern is the single strongest circumstantial argument that oestrogen sits somewhere in the causal chain.

Does body weight cause cellulite?

No. Body weight modifies how visible cellulite is; it does not create the underlying architecture. Lean women get cellulite, and losing weight can sometimes make it look worse by adding skin laxity to the same tethered septae.

The evidence points both ways at once, which is why the question confuses people. A 2008 Brazilian study — whose actual subject was lumbar posture, not cellulite prevalence — graded cellulite on a 1–4 scale across six body sites in 50 women with a mean BMI of 20.7 ± 1.9. It was a lean cohort in which cellulite was simply present and gradeable; it never reported how many of the 50 had it. On the other side, a 2015 correlation study found positive associations between BMI and clinical cellulite scores, and the cadaver biomechanics work concluded that female sex plus higher BMI produced taller superficial fat lobules and the greatest risk.

The reconciliation is straightforward. The septal architecture is the substrate; fat volume is the amplifier. More fat pushing against the same fixed tethers makes deeper dimples. Less fat makes shallower ones, unless the skin envelope has loosened, in which case appearance can worsen. The American Academy of Dermatology puts it plainly: “If you develop loose skin while losing weight, however, cellulite can become more noticeable.”

What roles do genetics, hormones, age and circulation play?

Of the factors blamed for cellulite, sex is the dominant one, and it acts through septal architecture. Age contributes through dermal thinning, which weakens the containment layer over the fat. Genetics and race appear in the review literature as risk factors but with thin supporting data. Circulation is a plausible modifier of septal fibrosis over years, not a proven cause.

Ageing is the best-quantified of these. In the cadaver series, increasing age was significantly associated with decreasing dermal thickness independent of sex (odds ratio 0.997, 95% CI 0.996–0.998, p < 0.0001). A thinner dermis is a weaker containment layer, so the same fat protrudes further. Because dermal thinning happens to men too, and men rarely develop cellulite, age is best understood as an aggravator rather than a cause.

Oestrogen’s role is inferred, not demonstrated by intervention trials. Reviews report that high-oestrogen states — pregnancy, nursing, long-term oral contraceptive use, hormone replacement therapy — appear to worsen the appearance of cellulite. That is observational and consistent with the androgen-deficiency pattern in men. Every source for this is observational. No trial has modulated oestrogen and measured what happened to cellulite grade.

Genetics is the weakest-evidenced item on the list. Reviews list genetics as a risk factor without naming a mechanism. No candidate gene, twin study or heritability estimate for cellulite appeared in any of the twenty-two sources read for this article, and none of those reviews cites one. That is not proof none exists; it is a statement that the reviews asserting heritability do not show their working.

Which cellulite myths are wrong, and exactly why?

Five claims about cellulite are wrong, and each fails for a specific mechanical reason: it is not trapped toxins, it is not just fat, exercise cannot remove it, dry brushing has no evidence base at all, and creams cannot reach the structure that causes it. What follows takes them one at a time, with what the measurement actually showed.

“It’s trapped toxins.” No toxin has ever been identified or measured in cellulite tissue. Direct in vivo measurement of fat metabolism and blood flow found no difference between dimpled and smooth skin on the same thigh.

“It’s just fat.” Fat is the filling, not the structure. If it were just fat, men with the same body fat percentage would have the same dimpling, and they do not. The variable that changes between the sexes is the connective tissue, which is why men’s fat protrudes less at a given thickness.

“You can exercise it away.” Exercise builds muscle underneath, which changes the surface contour. It does not resect a collagen septum. In a randomised controlled trial of 45 women aged 18–32 (42 completed), the only significant reductions in Cellulite Severity Scale score after six sessions across three weeks were in the arm combining aerobic exercise with radial shockwave therapy: against control (p = 0.042) and against the exercise-only arm (p = 0.032). The published abstract reports no result at all for exercise alone versus control, so this trial cannot be used either to convict or to acquit exercise on its own. Three weeks is short and the sample small — the mechanism argument is doing the work here, not this trial.

“Dry brushing removes it.” A PubMed search for dry brushing or skin brushing in relation to cellulite returns no indexed studies at all. There is no evidence base to evaluate. Brushing produces transient erythema — reddening from surface blood flow — which can make skin look temporarily smoother and more even. That effect resolves within hours and does not touch tissue at septal depth.

“Creams dissolve it.” Topical caffeine and retinol have shown statistically significant improvements in small, short trials, and they act by plausible routes: caffeine inhibits phosphodiesterase and can locally dehydrate cells; retinol increases dermal thickness over months. Neither reaches or remodels a 2 mm collagen septum. The AAD notes that caffeine products need daily reapplication to maintain any effect and that retinol requires six months or longer before anything is visible. A 2015 PRISMA-guided systematic review of 67 cellulite treatment studies concluded that “no clear evidence of good efficacy could be identified in any of the evaluated cellulite treatments.” The same review noted the two exceptions it found: some evidence of potential benefit for acoustic wave therapy and for the 1440 nm Nd:YAG minimally invasive laser.

Does any cellulite treatment last?

No treatment permanently removes cellulite. That is not a cautious hedge; it is the consistent position of every source reviewed for this article. Improvement means a temporary reduction in the visible depth or number of depressions, usually scored on a scale, usually requiring maintenance.

The durability figures are sobering when you line them up. Mechanical vacuum-roller massage (endermologie) produces measurable improvement in observational studies, but the AAD states that “the cellulite tends to return within 1 month of stopping treatment.” Radiofrequency and laser results are described in the review literature as short-lived and requiring repeated sessions. Acoustic wave therapy has shown improvement in several studies, but the review literature notes that six to eight sessions are usually needed before any visible reduction, and that durability data beyond one year are lacking.

The longest durability belongs to procedures that cut the septae. Gabriel’s review reports vacuum-assisted subcision results as lasting beyond three years. A 55-woman trial, reported at review level rather than read in the original, found about 94% of patients still one point or more better on the Cellulite Severity Scale at twelve months. Manual subcision is far weaker: Bass and Kaminer report a retrospective series in which improvement persisted past two years in only 9.9% of patients — 23 of 232 — with the follow-up denominator unclear and late recurrence common.

Only interventions that physically cut or enzymatically release the septae produce results measured in years rather than weeks, and they are surgical procedures with real costs. In the studies Gabriel’s review summarises, manual subcision caused excessive elevation of the treated area in 15% of subjects, painful bruising lasting up to four months in 90%, and hemosiderin pigmentation lasting up to ten months in every patient; releasing the septa too superficially risks skin necrosis. These are physician procedures. None of them is available at a wellness studio, including this one.

The enzymatic route carries a further caution. The injectable collagenase product approved in the United States in 2020 returned no active listing in the FDA’s drug label or NDC directories when last checked, which is consistent with it no longer being marketed here. Treat it as unavailable unless the manufacturer or the FDA says otherwise.

Bass and Kaminer’s conclusion sits underneath all of it: “the lack of durable efficacy and inconsistency seen in clinical results suggest that dermal or adipose tissue changes are not the primary etiologies of cellulite.”

Where is the evidence thinnest?

The evidence is thinnest on prevalence, on measurement and on durability. Several honest gaps remain. There is no population-level prevalence study. Instrument measurements and clinical scores correlate poorly. Most published trials are small, short, unblinded, and frequently sponsored by the maker of the device being tested. Where a scale was used, it is often the unvalidated 1978 four-grade scale, which cannot detect small changes. And the two scales built to be most sensitive to change were developed by a team dominated by the marketer of one of the products they were used to score.

Regulatory status is narrower than the marketing around it implies. Cellulite devices reach the US market through the FDA’s 510(k) pathway — that is clearance, not approval, and it is a finding of substantial equivalence to an existing device, not a finding that the device works well or that results last. The Resonic rapid acoustic pulse device (product code GEX, most recent clearance K233804, Zeltiq Aesthetics, February 2024) and the Cellfina vacuum-assisted subcision system (product code OUP, most recent clearance K192185, Ulthera, October 2019) both hold current clearances. No topical cream marketed for cellulite is FDA-approved for it. Before any specific device is named anywhere on this site, its indication for use should be quoted verbatim from its own 510(k) summary rather than paraphrased from a review article.

When should you see a doctor rather than a studio?

Cellulite is symmetrical, stable over years, and not painful at rest, although affected skin is sometimes tender when pinched — that much is described in the literature and is not a warning sign on its own. It does not appear suddenly. When what you are seeing does not fit that description, the sensible order is a medical opinion first and a booking second.

One item on that list should not wait. New or one-sided dimpling, puckering or thickening of the skin over breast tissue needs prompt medical assessment, because a peau d’orange change over the breast is a recognised presentation of inflammatory breast cancer, and it is the one item here where delay carries real cost.

Otherwise, ask a physician before booking any body treatment if you notice new or one-sided swelling, skin that pits when pressed, dimpling that appeared over weeks rather than years, pain present without pinching, warmth, or redness. Lipoatrophy, generalised oedema and lymphoedema can all look like skin depressions and have been mistaken for cellulite; treating them as cellulite can make them worse.

Lipoedema deserves its own mention. It is a symmetrical, disproportionate build-up of fat in the legs and sometimes arms that spares the feet and hands, feels tender or painful to touch, bruises easily, and does not respond to dieting the way the rest of the body does. It is frequently mistaken for cellulite or for simple weight gain and is under-diagnosed for years. If your legs hurt, bruise readily, feel heavy, or are out of proportion to the rest of you, see a doctor before booking any body treatment.

One naming note that matters. Cellulite and cellulitis are unrelated. Cellulitis is a bacterial skin infection: a spreading area of hot, red, painful, swollen skin, often with fever, that needs same-day medical assessment and antibiotics. If that is what you are describing, do not book a body treatment — call a doctor.

Also seek medical advice first if you are pregnant, have a bleeding or clotting disorder, an active infection or inflammation in the area, an implanted electronic device, or a history of blood clots. For lymphatic drainage specifically, tell us first if you have heart failure, kidney failure, an untreated or active cancer, or any current or suspected deep vein thrombosis. Moving fluid deliberately around the body is not neutral in any of those situations, and the answer may be that we should not treat you at all.

Lana Sculpt Studio is a wellness and body contouring studio, not a medical provider, and nothing here is a diagnosis or a substitute for medical care.

The most useful thing to hold onto is the anatomy. Cellulite is a tethering pattern in connective tissue that most adult women have, that most men are structurally protected from, that gets more visible with fat volume and with age, and that no cream, brush or squat has any mechanical route to remove. Treatments that make a difference make a temporary, measurable difference to appearance. Anyone telling you otherwise is describing something the literature has not found.