Why GLP-1 Weight Loss Varies So Much From Person to Person

The −14.9% and −20.9% headline numbers from the pivotal trials are averages, and averages erase the people on either side of them. Some trial participants lost far more than the mean; a meaningful share lost far less, even at the full approved dose. Here's what the actual response distributions show, and what a 2026 genetic study adds to explaining why.

The trial averages hide a wide range of real outcomes

STEP-1's headline is a mean weight change of −14.9% on semaglutide 2.4 mg — but the trial's categorical breakdown tells a more complete story: 86.4% of the semaglutide group achieved a weight reduction of 5% or more, 69.1% reached 10% or more, and 50.5% reached 15% or more, at 68 weeks¹. Flip those numbers around: roughly 1 in 7 participants did not reach even a 5% loss on the full labeled dose, and half did not reach 15%. SURMOUNT-1 shows the same spread at every tirzepatide dose — at the top 15 mg dose, 91% achieved 5% or more loss, but only 57% reached 20% or more; at the 5 mg dose, 85% reached 5% or more². These are large, real average effects and a wide range of individual ones, on the exact same dose, in the exact same trial.

Part of that variation now has a traceable genetic basis

A 2026 genome-wide association study in *Nature*, conducted with self-reported outcomes from 27,885 people who took a GLP-1 receptor agonist, identified a missense variant in the GLP1R gene — the gene coding for the receptor these drugs act on — significantly associated with greater weight loss, worth an additional 0.76 kg of loss per copy of the effect allele³. The same study found variation in GLP1R and GIPR linked to who experiences nausea or vomiting on treatment, with the GIPR association specific to people taking tirzepatide³. This doesn't mean genetics explains everything — it means at least some of what looks like "the drug didn't work as well for me" or "I got hit harder by side effects" has a real, measurable biological basis, not just a willpower or dosing story.

A plateau isn't automatically "the drug stopped working"

Before assuming true non-response, a few ordinary explanations are worth ruling out first. Undertitration is common — if dose escalation stalled below the labeled maintenance dose because of side effects or a cautious schedule, the trial evidence behind your expected result doesn't fully apply yet. Inconsistent dosing — missed or delayed weekly doses — quietly erodes results in a way that's easy to underestimate. And coverage interruptions matter more than people expect: our piece on insurance coverage walks through real-world data showing patients who lose coverage and restart inconsistently end up with outcomes far below what continuous treatment produces. Only once those are ruled out does a genuine biological plateau become the likely explanation.

What a provider should actually do about a below-average response

A good provider treats a lower-than-expected response as a data point to act on, not a reason to keep billing the same plan unchanged. That can mean confirming you've actually reached the labeled maintenance dose, addressing side effects that are limiting titration, or having an honest conversation about switching molecules — the trial evidence in semaglutide vs. tirzepatide shows the dual-mechanism drug produces larger average reductions, which is one legitimate option for someone who has plateaued on the single-mechanism drug at a fully titrated dose. What a provider should not do is treat a lower response as your failure rather than a clinical signal worth investigating.

Where this belongs in a provider's grade

Clinical oversight — 20% of the WeighScore — is precisely the dimension that separates a provider with a real plan for non-responders from one that only knows how to onboard new patients. If you're evaluating whether to continue, adjust, or stop, our weight-regain evidence is the companion piece worth reading before making that call, since stopping and restarting inconsistently carries its own well-documented cost.

The honest bottom line

The average trial result is real, but it was never a promise for any individual — the same studies that produced it show a wide range of outcomes on the identical dose, and a 2026 genetic study is starting to explain part of why. A plateau is worth investigating, not just accepting or panicking over. Compare graded providers on clinical oversight and dosing flexibility in our reviews and comparisons A licensed clinician who knows your full history and titration record should be the one interpreting your specific response.