How Deoxycholic Acid Works: Adipocytolysis, Inflammatory Clearance, and Why Swelling Is the Treatment

The single most common phone call after a submental treatment is about swelling. The deoxycholic acid mechanism of action is the reason that call happens, and it is also the reason the call is reassuring rather than alarming: the swelling is not a side effect sitting alongside the result. It is the second half of the result. Adipocytolysis destroys the fat cell; an inflammatory response clears what is left. Remove the inflammation and you remove the clearance.

That is the sentence I want every injector to be able to say out loud before they draw up the first vial. Not because it sounds good in a consultation — although it does — but because it determines dosing logic, injection plane, retreatment timing, and how you read a face at four weeks. This piece is the pharmacology underneath all four.

Maritza Mejia, FNP — family nurse practitioner; faculty member, Empire Medical Training; founder of Long Island Beauty Bar, New York.

What the molecule actually is

Deoxycholic acid is an endogenous secondary bile acid. Your patient already makes it. In normal physiology it sits in the bile acid pool and helps solubilise dietary fat in the gut so that it can be broken down and absorbed. The injectable formulation (ATX-101, marketed as KYBELLA in the United States) is a synthetically derived version supplied at 10 mg/mL in a 2 mL single-patient-use vial — 20 mg per vial.

It is approved for improvement in the appearance of moderate to severe convexity or fullness associated with submental fat in adults. The label carries an explicit limitation of use: safe and effective use for subcutaneous fat outside the submental region has not been established and is not recommended.

What matters pharmacologically is the word the label uses to classify it. The FDA label's mechanism of action section reads, in full: "KYBELLA injection is a cytolytic drug, which when injected into tissue physically destroys the cell membrane causing lysis."

Read that again and notice what is absent. There is no receptor. There is no enzyme. There is no metabolic pathway that a patient's physiology can be more or less responsive to. This is a detergent acting physically on a lipid bilayer. That single fact drives almost every practical rule that follows.

Why fat and not everything else

If deoxycholic acid is a detergent that disrupts any cell membrane it contacts, the obvious question is why injecting it into a neck does not simply dissolve the neck.

Thuangtong and colleagues addressed this directly in Dermatologic Surgery in 2010 (PMID 20482723). In vitro, deoxycholate lysed every cell type they tested across the concentration range studied. In vivo, injected into fat tissue, it killed adipocytes while other cell types appeared comparatively spared. The explanation they proposed is the one that has held: physiological concentrations of albumin, and protein-rich tissues generally, reduce deoxycholate's ability to lyse cells. Adipose tissue is relatively deficient in cell-associated protein and interstitial albumin. It cannot buffer the detergent. Muscle, gland and dermis, being protein-rich, partially can.

So the selectivity is real, but it is relative, local and concentration-dependent — not a property of the molecule that protects the patient on your behalf. Three chairside consequences follow immediately:

1. Depth is protection. Place the drug in protein-poor subcutaneous fat and selectivity works for you. Place it in the dermis and it does not. The label warns that injections that are too superficial may result in skin ulceration and necrosis.
2. Proximity is risk. The label instructs that the drug should not be injected into or in close proximity (1–1.5 cm) to salivary glands, lymph nodes and muscles. Those tissues are better buffered than fat, but "better buffered" is not "immune," and the inflammatory field that follows lysis is not selective at all.
3. Concentration, not enthusiasm, drives effect. Because this is a contact phenomenon between drug and membrane, the meaningful dosing unit is drug per unit area of fat, not millilitres per patient. That is why the label expresses dosing as an area-adjusted dose of 2 mg/cm² rather than as a total volume.

The histology, day by day

The most useful mechanistic dataset in this drug's file is a Phase 1 open-label study by Walker and colleagues, published in Dermatologic Surgery in 2020 (PMID 30883481). Fourteen adults received injections into abdominal fat at varying concentrations (0.5%, 1.0%, 2.0%, 4.0%), volumes (0.2 or 0.4 mL) and spacings (0.7, 1.0, 1.5 cm) at set intervals before a scheduled abdominoplasty. The tissue was then excised and examined.

Three findings deserve to be memorised.

Every injection paradigm produced changes confined to the subcutaneous layer. Not the dermis above it, not the fascia below it — the subcutaneous layer. When the drug is placed correctly, the injury it produces stays where it was placed.

The changes were more prominent at higher concentrations, and were independent of volume and spacing. Within the range tested, the strength of the solution in contact with the tissue drove the histological effect; how much of it you put in one hole and how far apart the holes were did not. This is the histological correlate of area-adjusted dosing.

The timeline is stereotyped:

Notice that the drug's own action — the detergent step — is essentially over within hours. Everything from Day 1 onward is the patient's tissue doing the rest of the work.

Why swelling is the treatment, not the complication

Humphrey and colleagues laid this out as a narrative review in Plastic and Reconstructive Surgery — Global Open in 2022 (PMID 37073386), and it is the clearest published statement of the principle. When injected into subcutaneous fat, deoxycholic acid physically disrupts adipocyte membranes, producing local adipocytolysis and cell death, followed by a mild local inflammatory reaction consisting of macrophage infiltration and fibroblast recruitment. By Day 28 that inflammation has largely resolved, leaving fibrotic septal thickening, neovascularisation and atrophy of fat lobules.

Put the two halves together and the logic is unavoidable. Lysis releases the contents of the adipocyte into the interstitium. Free lipid is not cleared by diffusion; it is emulsified and phagocytosed. Macrophages are the clearance mechanism. Macrophage infiltration produces oedema. Therefore oedema is the visible signature of the clearance step.

The pivotal trial numbers make the same point in a different language. Across the two Phase 3 trials supporting approval (513 treated subjects, 506 placebo), injection-site oedema/swelling occurred in 87% of treated subjects versus 43% of placebo. Pain 70% versus 32%. Numbness 66% versus 6%. Induration 23% versus 3%. Hematoma/bruising was 72% versus 70% — essentially a needle effect rather than a drug effect. A drug whose mechanism is lytic inflammation should produce inflammation in the overwhelming majority of the people it works in, and it does.

I teach this as a one-line reframe: swelling is the treatment, not the complication. A patient who swells is a patient whose macrophages are clearing lysed fat. A patient who does not swell at all is a patient I want to re-examine, because the most likely explanations are that I did not deposit into fat, or that there was less fat there than I thought.

That is not a licence to ignore severity. Oedema that is asymmetric, progressive after the first week, or accompanied by an asymmetric smile or difficulty swallowing is a different conversation and belongs in the adverse-event pathway, not the reassurance pathway.

The part of the result that is not fat loss

The Day 28 histology — septal thickening, neovascularisation, lobule atrophy — is the most under-taught part of this drug.

Some of what your patient sees at three months is not missing fat. It is a fibrotic remodelling of the fibrous septae that run through the subcutaneous compartment. That thickened septal network behaves like a tightened scaffold under the skin. It contributes to contour, and it explains a clinical observation that otherwise makes no sense: repeat treatments are usually less painful and less swollen than the first one.

Humphrey's review attributes that attenuation to a combination of factors — less target tissue remaining, which permits lower doses and injection volumes; persistent numbness from prior sessions; and greater tissue integrity from the thickened fibrous septa. All three are consequences of having already treated, not signs of tolerance in the pharmacological sense.

It also gives you something concrete to say when a patient asks whether the second session will be "as bad." It usually is not, and there is a structural reason.

The systemic picture: why this is not a weight loss treatment

I say this in every consultation and I want the reasoning behind it in every injector's head: this treats contour, not weight.

The pharmacokinetics are unambiguous. After a 100 mg dose — the maximum single treatment — deoxycholic acid is rapidly absorbed, reaching a median peak at 18 minutes with a mean Cmax of 1024 ± 304 ng/mL. That is 3.2-fold higher than baseline endogenous levels. Mean AUC0–24 is 1.6-fold higher than endogenous exposure. Plasma levels return to the endogenous range within 24 hours. The drug is 98% protein bound, is not metabolised to any significant extent under normal conditions, rejoins the endogenous bile acid pool and is excreted in faeces. No accumulation is expected at the labelled treatment frequency.

In other words: a maximal treatment produces a transient blip in a bile acid the patient already circulates, which is gone within a day. There is no systemic lipolytic effect. There is no metabolic mechanism by which this reduces body weight. Whatever changes in the submental region is a local tissue event in a defined compartment.

When a patient hears "dissolves fat" and imagines a systemic fat-dissolving injection, that misunderstanding is the origin of most disappointment. Correct it in the consultation, not at the four-week review.

What the evidence supports — and what it does not

The two pivotal Phase 3 trials enrolled adults aged 19–65 with BMI ≤ 40 who received up to six treatments at least one month apart. Assessed 12 weeks after the final treatment:

• ≥1-grade improvement on the clinician-reported submental fat rating scale: 66.5%–70.0% of treated subjects versus 18.6%–22.2% on placebo.
• ≥2-grade improvement: 13.4%–18.6% versus less than 0.1%–3.0%.
• MRI-measured ≥10% reduction in submental fat: 43% versus 5%.

That is a drug that reliably moves patients one grade and rarely moves them two. Long-term follow-up of REFINE responders (Humphrey et al., Aesthetic Surgery Journal, 2021; PMID 33617632) found that maintenance of one-grade response in the treated arm was 86.4% at Year 1, 90.6% at Year 2 and 82.4% at Year 3, with no new safety signals — so the result that is achieved is durable. Adipocytolysis destroys cells rather than emptying them, which is the mechanistic reason to expect durability.

What the evidence does not support is a dramatic single-session transformation, treatment of fat outside the submental region under the approved label, or any expectation of skin tightening as a primary outcome. A 12-month open-label study (Beer et al., Journal of Drugs in Dermatology, 2019; PMID 31524342) reported that at 12 months, 82.9% of subjects had unchanged skin laxity and 10.1% had improved laxity relative to 12 weeks after last treatment. Laxity mostly holds. It does not reliably get better, and that is a selection problem, not a dosing problem.

What this changes on Monday morning

1. Plane before dose. Selectivity is bought by depositing into protein-poor subcutaneous fat. Superficial placement forfeits it and risks ulceration and necrosis. Pinch, and inject into what you pinched.
2. Think density, not volume. A cytolytic contact mechanism plus histology that tracked concentration rather than volume or spacing is why the label's dose is expressed per square centimetre.
3. Consent to the inflammation specifically. Not "there may be some swelling." Name the mechanism, name the 87% figure, and tell the patient that the swelling is the drug working.
4. Stop reading results through oedema. Day 1–7 is the inflammatory phase. Anything you judge in that window is not the result.
5. Expect gradualism and plan for more than one session. Both the mechanism and the trial data say the same thing.
6. Do not sell it as weight loss. The pharmacokinetics will not support the claim, and the patient who wanted weight loss will be unhappy at a result that was technically excellent.

These figures reflect Maritza Mejia's clinical practice as taught in Empire Medical Training's hands-on curriculum, together with the current FDA prescribing information and published literature cited above. Technique is learned under supervision; this article is educational and is not a substitute for training.

Injectors who want to work through mechanism, patient selection, mapping and injection technique on live patients under faculty supervision can find that in Empire's Kybella and Mesotherapy Training, and the wider regional approach in the Neck & Hands Rejuvenation Master Course. If your gap is regional anatomy rather than the drug, Anatomical Based Aesthetics Training is the better starting point. For how a cytolytic agent differs from a collagen-stimulating one, see biostimulator injections versus fillers.

Related guides in this cluster

Part of Submental and Lower Face Contouring.

Clinical GuideThe Submental Pinch Test: Distinguishing Fat from Skin Laxity Before You Open a VialThe submental pinch test separates treatable fat from skin laxity. Maritza Mejia's thumb-and-index technique, thickness bands, and the Clinical GuideSubmental Anatomy for Deoxycholic Acid: Compartments, the Submandibular Extension and the Marginal Mandibular NerveThe safety anatomy of the neck for injectors — preplatysmal versus subplatysmal fat, the submandibular extension, and where the margina Clinical GuideDynamic Neck Assessment: Resting, Smiling, Flexion and ExtensionA four-position neck examination for injectors — what rest, animation, flexion and extension each reveal that a static pinch cannot, an