Masseter Botox: 5.24 Millimeters of Jawline, Zero for Your TMJ

Twenty Years Off-Label, One Press Release

On August 4, 2026, AbbVie announced that the FDA had accepted a supplemental Biologics License Application for BOTOX Cosmetic in "marked to very marked masseter muscle prominence." If it clears, it will be the first neurotoxin approved in the United States for the lower face, and the fifth aesthetic indication for a drug that started with frown lines in 2002.¹ No decision date was disclosed. Nothing about dose, responder rates, or duration was disclosed either. What was disclosed was that two US Phase 3 trials met their primary endpoints, and that roughly twice as many treated patients said they were satisfied compared with placebo.^1,2

Here is what makes this filing interesting to me. Injectors have been putting botulinum toxin into the masseter for cosmetic slimming since the mid-1990s, and into painful jaw muscles for even longer. Every one of those injections in the US has been off-label. The procedure is so mainstream that the phrase "masseter Botox" has its own subculture of before-and-after videos, most of them promising a "snatched" jawline, and a startling number of them promising relief from clenching, grinding, and "TMJ." A drug that has been used this way for decades, by this many people, should have a mountain of evidence behind it.

It doesn't. Or rather, it has a very specific kind of evidence, built almost entirely by the company that sells the drug, in a population that looks nothing like the people watching those videos, answering a question that is narrower than the one they're asking. That gap between the label and the hashtag is what this issue is about.

Why a Chewing Muscle Became a Beauty Problem

The masseter is the thick, rectangular muscle you can feel bulging at the back of your jaw when you clench your teeth. It runs from the cheekbone down to the angle of the mandible and it is, pound for pound, one of the strongest muscles in the body. Like any skeletal muscle, it hypertrophies with work. People who clench, grind, chew gum for hours, or simply inherit a large muscle belly develop a squared, widened lower face. Historically this was described as "benign masseteric hypertrophy" and treated, when it was treated at all, with surgery: partial resection of the muscle or shaving of the mandibular angle.⁸

Botulinum toxin changed that in the 1990s by offering a chemical alternative. The toxin blocks the release of acetylcholine at the neuromuscular junction, which means the nerve can fire but the muscle fibers don't receive the signal to contract. A partially paralyzed muscle atrophies, in the same way an arm in a cast does. Ultrasound studies show masseter thickness falling by roughly a quarter to a third at the three-month mark after a single treatment, with visible narrowing of the lower face beginning at two to four weeks and peaking around day 90.¹⁰ Then the nerve terminals sprout, the signal returns, and the muscle rebuilds over six to nine months.^3,10

That mechanism is elegant, well understood, and important to hold onto, because it defines who this works for. Botulinum toxin shrinks muscle. It does nothing to fat, nothing to a flared mandibular angle, nothing to a prominent parotid gland, and nothing to skin that has lost its elasticity. If your lower face is wide for any of those reasons, the toxin has no target. Allergan's own trial designers knew this. Their exclusion criteria read like a list of everyone who might buy the treatment based on a video: excess lower-facial fat, jowling, lax skin, parotid prominence, facial asymmetry, and, notably, any history or sign of temporomandibular disorder.²

Three Placebo-Controlled Trials. All Allergan. All Positive.

The cleanest way to describe the evidence base before 2020 is to quote the Cochrane Collaboration, which set out in 2013 to systematically review botulinum toxin for masseter hypertrophy and found zero randomized or controlled trials eligible for inclusion.⁸ Two decades of clinical use, and nothing that met the bar. The literature was case series, mostly from South Korea, Taiwan, and China, mostly uncontrolled, mostly small.

Allergan then built the evidence itself. Three placebo-controlled trials followed, and I want to walk through each of them, because they are genuinely well designed and they are also the only controlled data that exists.

Then come the two US Phase 3 trials that underpin the FDA filing. M21-416 enrolled 202 patients and M21-417 enrolled 207, both at about 20 US sites, both 3:1 toxin to placebo, both triple-blind, both completed in November 2025.² Both met the primary endpoint. What AbbVie reported was the p-values: 0.0046 and 0.0014.¹ I'll note, carefully, that those are respectable but considerably less emphatic than the p<0.0001 from the multiregional trial. That could reflect smaller effect sizes in a non-Asian population, more conservative trial sizing, or nothing at all. Until the data are published, nobody outside AbbVie knows, and that includes the influencers already treating the filing as a vindication.

Who Was Actually in the Room

Every one of those trials enrolled people with bilateral grade 4 or 5 masseter prominence on a five-point scale, confirmed by both the investigator and the patient, with the two ratings required to match. Body mass index had to be 30 or below. Nobody with jowls, lax skin, a fat-heavy lower face, or jaw joint symptoms got in.² This is good trial design; it isolates the population in whom the mechanism should work. It is also a population that is, by construction, a small slice of the people who will be sold this treatment once "FDA-approved" appears in clinic marketing.

The demographics matter too. The pivotal multiregional trial was 87% Asian and 82% specifically Chinese, with a mean age of 31.³ That reflects where the demand and the historical literature have been concentrated, and it is entirely reasonable as a starting point. But it means the two US trials are the first meaningful controlled dataset in a broader population, and their demographics have not been published. We do not yet know how the drug performs in a 45-year-old with early jowling, because the trials were designed to keep her out.

And then there is the satisfaction number. In the best-case population, treated at the higher dose, with a validated scale and blinded raters, 53.9% of patients said they were satisfied at three months.³ That is a real, placebo-beating result. It is also a coin flip. The 5.24-millimeter width reduction is measurable, and it is visible in standardized photography. It is also about a fifth of an inch across the whole face. If the mental image you carry into the clinic came from a filtered video, the data suggest a roughly even chance you'll walk out underwhelmed.

The Claim the Label Will Never Support

This is the part I most want readers to take away. A substantial share of masseter toxin marketing is not about jawlines at all. It's about clenching, grinding, morning headaches, and "TMJ," a catch-all that usually means temporomandibular disorder, or TMD. The pitch is that relaxing the masseter relieves the pain, and that the cosmetic slimming is a bonus. Some clinics run it the other way around. Either way, the implication is that the two benefits come as a package.

They do not, and the evidence for the pain half is much worse than the evidence for the cosmetic half. The most rigorous recent synthesis, a 2024 meta-analysis of 14 randomized trials in PLOS One, concluded that botulinum toxin was "not significantly better than placebo" for TMD pain at one, three, and six months, and found no difference in maximum mouth opening, bruxism events, or maximum bite force.¹³ A 2023 meta-analysis in Aesthetic Plastic Surgery pooling ten bruxism studies found short-term pain reductions that scaled with dose, but no difference in maximal biting force versus saline (p=0.50).¹⁴

It gets worse. A 2025 meta-epidemiologic study in BMC Medical Research Methodology examined the bruxism trial literature specifically for "spin," meaning language that presents results more favorably than the data warrant. The authors found spin in 59.4% of the trials overall and in 87.5% of the conclusions in the main text.¹⁶ The commonest tactics were inadequate extrapolation and overstated clinical implications. In other words, even the trials that exist tend to say more than they show.

Now connect this to the FDA filing. The pivotal aesthetic trials excluded anyone with a history or sign of TMD.² An approval for masseter prominence will therefore provide precisely zero evidentiary support for treating jaw pain, and Botox will remain unapproved for TMD and bruxism. Any clinic that markets the new indication as validation of "Botox for TMJ" is misrepresenting what the label says. I expect a great many of them to do it anyway.

What Repeated Paralysis Does to the Jaw Underneath

Bone is not inert. It remodels continuously in response to mechanical load, and the mandible is loaded almost entirely by the masticatory muscles. Take the load away and the bone responds the way an astronaut's femur does in microgravity. This is not a theoretical concern; it has been demonstrated repeatedly in animals. Adult rats given botulinum toxin in the masseter and temporalis lost about 20% of alveolar bone volume and 35% of condylar bone volume.¹⁹ A 2019 review in Toxins compiled mouse, rat, and rabbit data showing consistent bone loss at the mandibular condyle, on the order of 10 to 11% by micro-CT.²⁰ Repeated injections in rabbits produced cumulative damage.²⁰

Humans are not rats, doses differ enormously, and a single cosmetic treatment is not the same as chronic paralysis. But the human signal is not zero either.

AbbVie's response was to run its own imaging. A prespecified CT substudy of 123 patients from the Phase 2 trial, treated with 48 to 96 units once or twice over 12 months, found "no clinically significant changes" in mandibular bone density at the condyle, dentoalveolus, or ramus.⁶ A companion analysis of mandibular morphology in 187 patients found no change in bigonial width, cortical thickness, or gonial angle, and 94% had no qualitative change at the condyle.⁷ These are reassuring, and they are better data than the animal work for the question that matters: one or two treatments, a year of follow-up, standard cosmetic doses.

The problem is the phrase "one or two treatments, a year of follow-up." That is not how this drug is used. Cosmetic patients re-dose every four to six months, often for years. In the company's own morphology paper, the authors concede that "late-emerging effects beyond 12 months cannot be excluded."⁷ Both studies were Allergan-funded and AbbVie co-authored. I do not think that makes them wrong. I think it makes them the floor of what we know, not the ceiling, and the ceiling is where the real-world exposure lives.

One more independent data point. A 2024 randomized, triple-blinded trial in Scientific Reports gave 26 women a single masseter treatment, then randomized them at three months to a saline booster or a second toxin dose. After a single injection, muscle thickness, electrical activity, and chewing performance all recovered by three months. After the second injection, thickness had not recovered at six months, electrical activity was still suppressed, and chewing performance remained impaired throughout.⁹ The authors' conclusion was blunt: the toxin works for hypertrophy, but "multiple injections cause functional adverse effects." That is the first controlled evidence that the repeat-dosing regimen everyone actually uses behaves differently from the single-dose regimen everyone studies.

Side Effects, Costs, and the Subscription You Didn't Sign Up For

The short-term side effect profile is well characterized and mostly tolerable. The largest real-world series, 2,036 masseter treatments at a single Taiwanese clinic, reported temporary loss of chewing force after 30% of sessions, bruising after 2.5%, headache after 0.58%, paradoxical bulging of the untreated superficial fibers after 0.49%, a sunken cheek after 0.44%, and a limited smile after 0.15%.¹¹ The smile problem happens when toxin diffuses forward into the risorius or zygomaticus muscles; it is rare, it is temporary, and it is the complication patients fear most. A 2018 literature review spanning 1994 to 2018 added "worsened jowls or sagging" to the catalogue, without controlled incidence data, and noted that most adverse events appear at two to four weeks and resolve within twelve.¹²

Allergan's Phase 3 numbers were lower: mastication disorder in 4.2%, paradoxical bulging in 2.1%, and facial paresis in 0.7%.³ Eight of 283 treated patients (2.8%) versus one of 93 on placebo had events flagged as "possible distant spread of toxin," including one case each of Bell's palsy, shortness of breath, and a visual accommodation disorder, all adjudicated as local or unrelated.³ Botox carries a boxed warning for distant spread, and it will continue to.

Then there is the cost, which is where the "subscription" framing becomes literal. A typical US session runs $400 to $1,200 depending on units and geography, with 20 to 30 units per side for cosmetic use and more for clenching.¹ Effects last four to six months in practice, sometimes longer with cumulative treatment.^3,10 Call it $1,200 to $2,400 per year, indefinitely, because the muscle rebuilds the moment you stop. And with each round, you are adding to a cumulative-exposure denominator that no controlled study has yet measured.

Strong for the Muscle. Silent on the Bone. Nothing for the Joint.