Key takeaways
- GLP-1 therapy has two dose-response curves: the benefit curve flattens as receptor occupancy saturates, and the side-effect curve does not.
- Three things flatten the benefit curve - receptor saturation, metabolic counter-regulation against the deficit, and behaviour becoming the rate-limiter at the top.
- Titration schedules exist because the gut adapts to slowed emptying over weeks; a step interval is a minimum, not an appointment.
- Judge a step up on appetite, rate of change, whether side effects are settling, and whether intake quality held - not on the scale alone.
- The lowest dose that does the job is the goal, because tolerability is an efficacy issue routed through adherence.
The assumption behind most dose escalation is that the medication has one curve: put more in, get more out. GLP-1 therapy has two curves, and they do not have the same shape. The benefit curve rises steeply at first and then flattens. The side-effect curve does not flatten. Everything that matters about dosing follows from the gap between those two lines — and the gap is different for every person, which is why "go to the top of the range" is a schedule, not a decision.
Two curves, not one
A dose-response curve describes how much effect you get for a given amount of drug. For a GLP-1 receptor agonist, the effect people care about — reduced appetite, earlier satiety, less food preoccupation — depends on how much of the receptor population is occupied and signalling. That relationship is not a straight line. Receptor systems saturate: once most receptors in the relevant brain and gut tissue are engaged for most of the dosing interval, more drug adds less signal, because there is less unoccupied receptor left to engage.
Modelling of individual weight-loss trajectories across the dose range shows exactly this shape: response builds as the dose is titrated upward and the curve flattens at the top (Strathe et al., Diabetes Obes Metab 2023). The tolerability curve behaves differently. Nausea, constipation, reflux and fatigue are largely driven by gastric emptying and gut motility effects, and those keep scaling with exposure — they do not politely plateau at the same point the appetite effect does (Kunutsor et al., Drugs 2026).
So the last step up the ladder costs the most and buys the least — the whole argument in one sentence, and it holds for both molecules in common use.
Why the benefit curve flattens
Three things happen as the dose climbs, and they work against each other.
Receptor occupancy saturates. The relationship between drug concentration and receptor activation is hyperbolic, not linear. Early increments move occupancy a lot; late increments move it a little. That is standard pharmacology, not a quirk of this class — GLP-1 receptor biology covers where those receptors sit.
The body counter-regulates. Energy deficit triggers adaptation: appetite signalling pushes back, resting expenditure falls as mass falls, and spontaneous movement drops. A higher dose is pushing against a system defending itself. Why a GLP-1 stops working deals with the plateau this produces.
Behaviour becomes the rate-limiter. At sufficient dose, appetite is no longer the constraint. What food is available, what gets eaten, whether protein and resistance training are in place — those decide the outcome. Adding drug does not fix a food environment.
Why titration exists at all
Escalation schedules are not there to build up to a target as fast as tolerated. They exist because the gut adapts to slowed emptying over a few weeks, and stepping the dose gives that adaptation time to happen. Skip the steps and the nausea is worse than it needed to be at the same eventual exposure — which is the pattern seen when escalation is compressed (Wharton et al., Diabetes Obes Metab 2022).
The practical consequence: escalation should be tolerability-paced, not calendar-paced. A step interval is a minimum, not an appointment. Holding longer than the schedule suggests is a normal clinical decision, not a failure to progress. So is stepping back down.
What each step up actually buys and costs
| Early steps | Late steps | |
|---|---|---|
| Appetite effect | Large change from a small increment | Small additional change |
| GI side effects | Present, usually settle with adaptation | Rise, and are less likely to settle |
| Adherence risk | Low | Higher — this is where people stop |
| Main constraint on results | The medication | Food quality, protein, training |
| Right question | "Is this doing anything yet?" | "Is this doing enough more to be worth it?" |
How to tell where you are on your own curve
Nobody knows their curve in advance. You find it by reading four signals over the four weeks after a step up:
- Appetite, not weight. Weight is noisy week to week. The cleaner readout is whether meals are ending earlier and whether food preoccupation has dropped — see food noise for why this is the more reliable early signal.
- Rate of change. If the trend line was already moving at a rate you are content with, the step up was optional.
- Side-effect burden. Not just whether nausea exists, but whether it is fading week on week. Settling is adaptation. Not settling is the dose.
- Intake quality. If protein and vegetables are being crowded out because nothing is appealing, the dose is already past useful — the tissue you are losing has changed character.
A step up that produces no change in appetite but more nausea is information: you were already at or past the flat part of your curve. That is a reason to go back down, not to wait it out.
The lowest effective dose is a real goal
There is a persistent belief that a lower dose means a lower-quality result. On a flattening curve, it can mean the opposite. The dose you can take consistently for a long time beats a higher dose you abandon in month four, because this is a chronic-condition medication and the outcome depends on staying on it. Tolerability is not a comfort issue; it is an efficacy issue routed through adherence.
Lower exposure also tends to preserve the ability to eat properly, which matters more than it sounds. Rapid weight loss with suppressed intake takes lean tissue with it. The defence against that is protein and resistance training, and both require actually being able to eat and train — see the muscle preservation playbook and how resistance training changes what a GLP-1 does.
Lower-dose use for goals other than maximum loss
Not every goal wants the steepest possible trajectory. Some people are using this class for body recomposition rather than large-scale loss, some for maintenance after a goal has been reached, and some primarily for appetite regulation or glucose handling. In each case the target is a stable, tolerable signal rather than the strongest one available — a physician's judgement about the individual, not a protocol to assemble from an article.
The clinical pearl: escalate only if the last step earned it. Someone who is losing steadily, eating properly and training at a moderate dose has no clinical reason to climb — the maximum dose exists for people whose response requires it, not as a destination everyone is meant to reach. The right question at every step is not "what dose am I due for?" but "what did the last increase actually change?"
What to expect over the first several months
The first weeks are mostly about the gut adjusting, not about weight — appetite changes usually arrive before anything meaningful shows on the scale. Through the middle of the titration the rate of change is typically at its fastest. Later, most people reach a point where the next step no longer changes how they eat. That point is the top of their useful range, and it is frequently below the top of the licensed range. Loss then continues on the strength of the deficit already in place, and eventually slows as the body reaches a new equilibrium — normal, not failure.
If none of the four signals are moving at a dose you tolerate well, that is a different conversation from "increase it": non-response has its own differential and deserves an evaluation rather than another step. To work out whether this class fits your situation at all, the 60-second assessment is the starting point, and any dose decision belongs with the prescribing physician.
Bottom line
GLP-1 dose-response is non-linear: the benefit curve flattens as receptor occupancy saturates and the body counter-regulates, while the side-effect curve keeps climbing. The highest dose is therefore not automatically the best dose, and for many people the useful ceiling sits mid-range. Escalate on tolerability and evidence of effect rather than the calendar, judge each step by what it actually changed, and treat the lowest dose that does the job as the goal rather than a compromise — because the dose you can stay on is the one that produces the result.
Educational content, not medical advice. Laboratory interpretation and any treatment decision are made by a licensed physician after individual evaluation. Individual results vary.
