GLP-1 Dose-Response: Why Lower Doses Do Less

One claim sits underneath most of the microdosing pitch: that you can shrink the dose and keep most of the benefit. The dose-response evidence — trials that deliberately compared different doses of the same drug — says otherwise. The relationship between dose and effect for GLP-1 drugs is graded: more drug, more effect; less drug, less effect. This page walks through the numbers.

The clearest evidence: semaglutide dose-finding

The single best source on this question is a phase 2 dose-ranging trial of semaglutide for weight loss ¹. Because it tested several doses head-to-head against placebo, it shows the dose-response curve directly. The mean weight-loss results:

• •Placebo: about −2.3%
• •0.05 mg: about −6.0%
• •0.1 mg: about −8.6%
• •0.2 mg: about −11.6%
• •0.3 mg: about −11.2%
• •0.4 mg: about −13.8%

The trend is hard to miss. The smallest dose tested (0.05 mg) produced roughly 6% weight loss; the largest (0.4 mg) roughly 14%. Effect size climbed as the dose climbed. A "microdose" — by definition a small fraction of a standard dose — sits at the bottom of this curve, in the territory of the smaller effects. This is the central reason the honest expectation for microdosing is less benefit, not equal benefit. We apply this curve specifically to the sub-therapeutic question in microdosing semaglutide: does a sub-therapeutic dose do anything?.

It's worth being precise about what a curve like this does and doesn't say. There are signs of flattening near the top — the 0.3 mg and 0.4 mg results are close, which is why drug developers eventually settle on a standard maintenance dose rather than pushing ever higher. But that flattening is at the high end of the curve, not the low end. Down where a microdose lives, the slope is steep: each step down in dose costs a visible chunk of effect. Nothing in this data suggests you can drop to a fraction of the dose and stay near the plateau.

The tiered trials say the same thing

You don't have to rely on a single dose-finding study; the large pivotal trials were built with dose tiers, and they tell the same story.

Tirzepatide's obesity trial reported roughly 15–21% weight loss across its 5, 10, and 15 mg tiers — with the higher tiers producing greater loss ². Its diabetes program showed the same gradient: in SURPASS-1, tirzepatide produced large, dose-dependent reductions in HbA1c and weight in type 2 diabetes ³. And in the head-to-head SURPASS-2 trial, tirzepatide's 5, 10, and 15 mg doses produced clearly dose-graded improvements in HbA1c and weight ⁴. Across drug, population, and outcome, the same pattern holds: higher dose, larger effect.

Why this happens: exposure-dependence

The reason lower doses do less isn't mysterious. These drugs work by occupying GLP-1 receptors enough to change appetite, gastric emptying, and glucose handling — and that effect depends on having enough drug present. When exposure drops, so does the effect, and when exposure falls away entirely the effect reverses. A maintenance trial showed this vividly: people who stopped semaglutide regained weight, while those who continued held their loss ⁵. The benefit tracks ongoing, adequate exposure.

For microdosing this is the crux. A microdose is a deliberate reduction in exposure. The dose-response and maintenance data together imply that a low enough dose may sit below the level needed to produce — let alone sustain — a meaningful effect.

It's also worth remembering that the standard doses weren't chosen arbitrarily. They are the doses that cleared the bar in large trials and earned regulatory approval, balancing effect against tolerability. A microdose is, in effect, a self-selected dose that has never been measured against that bar. The dose-response curve is the closest thing we have to a preview of how it would perform — and that preview points downward.

What this means for microdosing

Put the pieces together. The dose-finding data show effects shrinking as the dose falls. The tiered pivotal trials show the same gradient across two drugs. And the maintenance data show benefits depend on keeping exposure up. None of that supports the idea that a microdose delivers most of the benefit. The weight of the evidence points the other way: lower doses do less.

This doesn't prove a microdose does nothing — small doses can produce small effects — but it removes the foundation from the "same benefit, smaller dose" promise.

For the full context, read our pillar: Microdosing GLP-1 — what the evidence actually shows. You may also want does microdosing GLP-1 actually work and is compounded / microdosed GLP-1 safe, or see how this curve frames the tradeoff in low-dose vs full-dose GLP-1: what actually changes, and how these numbers map onto the schemes people share in our microdosing tirzepatide dose chart. For a condition-specific case, see does microdosing GLP-1 help PCOS, and if you're coming off the drug, how to taper off GLP-1 without regaining. The same dose-response logic explains why microdosing retatrutide, an unapproved molecule is unproven — its lowest studied dose, 1 mg, already produced real weight and liver-fat changes, with nothing tested below it. To compare providers, use the GLP-1 microdose rankings hub.