The two trials that supported approval, and every pooled analysis built on them, used the same safety method: record each side effect, record the date it started, record the date it stopped. That produces incidences and durations. It does not produce a curve, because nobody was scoring how swollen each person looked on each day.
So the honest answer to "what will day three look like" is that it has not been measured: What can be described is the biology, which is documented, and the durations, which are documented, and the first-session effect, which is documented. That is a less tidy answer than a numbered timeline, but the numbered timelines are not data.
Kybella is deoxycholic acid, a bile acid the body already makes. The FDA-approved label describes the whole mechanism in one sentence: it is a cytolytic drug that, injected into tissue, physically destroys the cell membrane and causes the cell to burst.
That is a detergent action, and in a test tube it is not fussy about what it dissolves. Laboratory work shows the reason it mostly spares other tissue is local chemistry rather than any built-in targeting: protein-rich and albumin-rich tissue neutralises the detergent, and fat, which is poor in both, cannot. "Selective for fat" is a real effect, but it comes from what surrounds the drug, not from the drug recognising fat cells.
Once fat cells rupture, their contents have to be cleared, and clearance is an inflammatory job. A Phase 1 study injected the drug into abdominal fat 1, 3, 7 or 28 days before people underwent a planned abdominoplasty, then examined the removed tissue. On day 1 there was fat-cell destruction, injury to small blood vessels and neutrophil inflammation. By day 3 the inflammation had eased but there was bleeding into the tissue and pooled lipid. At day 7 the destruction was still prominent, with lipid-filled scavenging cells at work and the damaged vessels repairing. At day 28 inflammation had largely resolved, leaving thickened fibrous partitions, new small blood vessels and shrunken fat lobules. All the changes stayed in the fat layer. The authors were employees of the developer, and this was abdominal skin rather than the neck, but it is the clearest picture available of what is going on under the swelling.
The label reports the two trials together: 513 people treated, 506 on placebo, followed for up to six months after their last treatment.
Injection-site reactions of any kind occurred in 96 per cent of treated patients. Swelling or oedema, reported as one combined term, occurred in 87 per cent, against 43 per cent on placebo. Placebo injections cause swelling too, which is worth holding onto: some of this is the needle and the volume, not the drug. But the gap between 87 and 43 is the drug.
The single most useful sentence in the label for this question is the one about duration: reactions lasting more than 30 days, and affecting more than 10 per cent of subjects, were numbness at 42 per cent, swelling or oedema at 20 per cent, pain at 16 per cent and firmness at 13 per cent. One in five people was still swollen more than a month after a session, and that is from the drug's own labelling, not from a critic.
The label stops there. It gives no median, no peak day and no average.
The medians people quote come from a post hoc pooled analysis of the same two trials, published in 2018. It reports, for the treated group: bruising median 9 days, pain 7 days, oedema 10 days, swelling 11 days, redness 3 days, firmness 28 days, lumps 23 days and numbness 43 days.
Two things have to be said about those numbers. The analysis was funded by the drug's developer and three of its authors were company employees, with a fourth from the acquiring company. And a post hoc analysis is a look back at data collected for another purpose, not a study designed to answer this question.
The same paper carries the sentence clinics quote most often: injection-site reactions, while common, "typically resolved within 14 days" without medication or procedures. That is the authors' summary of their own medians, and it sits alongside the label's finding that a fifth of people were still swollen past 30 days. Both are true. A median is the day by which half the episodes had ended.
An independent-looking check exists. A single-centre prospective series of 100 consecutive patients in private practice reported mean durations of 7.1 days for local oedema, 3.5 days for tenderness and 27.9 days for numbness. Its author also receives support from the manufacturer. The oedema figure is shorter than the trial median, which is what you would expect from a real-world group receiving fewer sessions than trial protocol allowed.
Assembling only sourced points, and treating them as what they are rather than as a curve:
Within the hour: The drug peaks in the bloodstream at a median of 18 minutes and is back in the body's normal bile-acid range within 24 hours. Whatever is happening after day one is happening in the tissue, not from drug still circulating.
Days 1 to 3: The peak inflammatory phase on histology. Redness had a median duration of 3 days in the trials, and tenderness a mean of 3.5 days in the real-world series. This is where firmness and fullness are most reported clinically, though no study scored the swelling itself day by day.
Around day 7: Histology shows scavenging cells clearing lipid and blood vessels repairing. Pain had a median of 7 days. The real-world series put mean oedema at 7.1 days.
Days 9 to 11: The trial medians for bruising, oedema and swelling all fall here.
Around day 14: The point the pooled analysis authors picked as the typical resolution of injection-site reactions.
Days 23 to 28: Lumps had a median of 23 days and firmness a median of 28 days, so the last thing to go is often texture rather than volume. Histology at day 28 shows inflammation resolved and the fibrous partitions in the fat thickened.
Past 30 days: 20 per cent still had swelling, 42 per cent still had numbness.
A median is not a curve: Half of all episodes ran longer than the numbers above, and the label's more-than-30-days figures are the proof of that. Anyone quoting you a fixed number of days is quoting a middle, not a promise.
Swelling is worst after the first session: some adverse event followed 94.4 per cent of first sessions and 46.4 per cent of sixth, and the downtime data run the same way. The session-by-session figures are on the Kybella sessions page.
If you are planning around this, plan around session one: Sessions have to be at least a month apart under the label, so there is time to see how your own first round goes before committing to the next.
Difficulty swallowing occurred in 2 per cent of treated patients in the trials, against under 1 per cent on placebo. The label is specific that it happened "in the setting of administration site reactions" — pain, swelling and firmness under the chin — and the pooled analysis links it to injection volume and post-injection swelling. It is a swelling effect, not a nerve injury. Cases resolved on their own, median 3 days, with a range of 1 to 81 days.
Two caveats matter. People with a current or past history of swallowing difficulty were excluded from the trials, and the label says to avoid the drug in them, so the 2 per cent figure comes from a screened population. And 2 per cent is not zero.
Difficulty swallowing or breathing is not something to monitor at home. It is an emergency, and the section below says so.
Swelling is also what people confuse with the one genuinely serious injection injury, marginal mandibular nerve paresis. That, and the numbness, lumps and hair loss, are covered on the Kybella recovery page.
Say if you have ever had trouble swallowing, even briefly and even long ago. The label says to avoid the drug in people with a current or prior history of it, and those people were excluded from the trials.
Ask how many millilitres are planned for your first session and whether the plan is to start smaller. The first session carries the most swelling, and the pooled data suggest volume and swelling travel together.
The rest of the pre-treatment checklist — bleeding risk, what the fullness is actually made of, and the infection rule — is on the Kybella recovery page.
Difficulty swallowing, or any sense that your throat or airway is closing
Difficulty breathing
Fever with spreading redness, warmth or pus from the treated area. These are signs of a spreading infection, and reported infections after this drug have included abscesses needing intravenous antibiotics and surgical drainage
Your smile becomes uneven, or you cannot move your lower lip normally on one side
Swelling, redness or pain is increasing after about day three instead of settling
An open sore, a blister, a dark or dusky patch, or any drainage appears at an injection site
Swelling is still obvious past four weeks, or a firm lump has not started to soften by then
Hair thins or disappears at an injection site
You feel faint or lightheaded during or after the injections
The swelling makes it uncomfortable to swallow but not frightening. Say so before the next session, because it is tied to injected volume