Botulinum Toxin for Frown Lines: Reading the Evidence Behind the Marketing

Botulinum toxin type A softens moderate to severe frown lines for most treated adults at roughly one month, and randomized placebo-controlled trials establish that effect cleanly and consistently across products. What the trials do not settle is the advertising. A glossy "responder rate" is only as meaningful as the cutoff a study chose, and the small chance of a temporary drooping eyelid almost never appears in the same font size as the before-and-after photo. Learn to read those two numbers and you have most of what separates the science from the sales pitch.

Key points#

• Against placebo, the benefit for glabellar (between-the-eyebrows) lines is large, reproducible, and holds across the marketed formulations.
• A "responder rate" is a fact about a chosen threshold, not a fixed property of the drug, so two honest studies can quote very different percentages.
• Temporary eyelid droop (blepharoptosis) runs in the low single digits, higher than placebo but neither zero nor common.
• Efficacy figures describe one treatment at about day 30 in carefully selected adults, not month four or years of repeat cycles.

First, decode the "responder rate"#

Start with the number the marketing leans on hardest, because it is the easiest to misread. A responder rate sounds like a property of the medicine, the way a boiling point is a property of water. It is not. It is a fact about where someone drew a line.

The pivotal trials graded frown lines on a short ordinal scale, usually running from 0 (none) to 3 (severe), and most used two co-primary endpoints rather than one: a trained investigator's rating of the crease at maximum frown, and the treated person's own global assessment of change. The FDA label for onabotulinumtoxinA (BOTOX Cosmetic) states the thresholds plainly. On the investigator scale, a "responder" had to reach a grade of 0 or 1 at maximum frown. On the subject scale, a responder had to report at least a +2 (moderate improvement) on a range from +4 to -4.

Change either threshold and the headline moves without a single new patient. Demanding a drop from severe all the way to "none or mild" is a stricter test than counting any one-grade improvement, and the stricter test yields a smaller percentage from identical data. So when two products quote different response numbers, part of that gap can be the definition, the timepoint, or whether the face was scored at rest or mid-frown, rather than anything about the molecule. When you compare two figures, ask what counted as a response, when it was measured, and who did the scoring.

What the pooled evidence actually shows#

Once the definitions are matched, the direction and size of the effect are not in dispute. A 2023 network meta-analysis in Aesthetic Plastic Surgery pooled randomized trials of five botulinum toxin type A formulations for glabellar lines and found that every one significantly outperformed placebo on the standard response endpoints, with broadly comparable rates of treatment-related side effects across products. A 2026 systematic review in Aesthetic Surgery Journal Open Forum applied the GRADE framework to three phase III placebo-controlled trials of incobotulinumtoxinA across the upper face and reported that the share of patients reaching at least a one-grade improvement at day 30 was far higher with active treatment than placebo in the glabella, forehead, and crow's-feet regions.

Two qualifiers belong beside those results. First, an eye-catching relative effect can be partly an artifact of arithmetic: when almost no one in the placebo group improves on their own, any ratio against that near-zero baseline balloons, so a risk ratio in the double digits looks more astonishing than the absolute benefit warrants. Second, the whole evidence base is anchored to a narrow window, a single treatment read at about a month in adults specifically selected for moderate to severe lines. It is not a promise about how your face looks at four months, or after several repeat cycles.

The side effect the ads skip: eyelid droop#

The local adverse event that deserves the most attention is blepharoptosis, a temporary drooping of the upper eyelid that happens when a little toxin reaches the muscle that lifts the lid. The registration data describe it in modest but nonzero terms. The onabotulinumtoxinA label reported blepharoptosis in about 3 percent of treated subjects versus none on placebo in the original trials, and in a repeat-injection study the rate was roughly 2 percent in the first cycle and 1 percent in the second. The incobotulinumtoxinA meta-analysis reported brow ptosis, a droop of the eyebrow that is distinct from the eyelid, as numerically higher with active treatment but not statistically separable from placebo, while serious adverse events were rare and none were judged treatment-related.

Put those figures side by side and the honest reading is a low single-digit, usually self-limited risk. A claim that a product is "complication-free" contradicts its own label; a claim that eyelid droop is frequent overstates it. The accurate summary is a small, real, generally transient chance that a reasonable person can weigh against a cosmetic benefit. How much that possibility matters is your judgment, and it belongs in a conversation with a qualified clinician who can examine your face, not in a blog post.

A short checklist for the next claim#

A handful of questions cut through most of the advertising.

• What was the responder definition, and at what timepoint?
• Was the comparison against placebo or against another active product?
• If a product is called "superior," is that a head-to-head trial or an indirect network estimate?
• Does the safety line quote the labeled side-effect range, blepharoptosis included, or does it fall silent there?

A claim may legally describe what a trial showed in the population studied; it may not legally promise a specific result for you. The evidence establishes that these drugs beat placebo for glabellar lines on defined endpoints with a modest, well-characterized risk profile. Past that line is where the data stop and the persuasion starts.