Ozempic Stopped Working? GLP-1 Plateaus Explained

Look at the weight curve in the STEP 1 trial and you will see something the headlines skipped: it flattens. Participants on semaglutide lost steadily for about a year, then the curve leveled off near 15% total loss and stayed there through week 68 (Wilding et al., 2021, New England Journal of Medicine, PMID 33567185). The plateau is not a malfunction. It is in the trial data of every weight-loss drug ever tested.

So when the scale stops moving in month nine and the Reddit thread fills with “Ozempic stopped working for me,” the more accurate sentence is usually: the drug is still working, and the system around it reached equilibrium.

Why every curve flattens

Three forces converge. First, intake drifts up. Appetite suppression is strongest early; over months, eating expands to fill the space the drug left, a few hundred calories at a time, mostly untracked. Self-report data shows how invisible this is: in a classic study of people who insisted they could not lose weight on very low calories, actual intake was 47% higher than reported and physical activity 51% lower (Lichtman et al., 1992, New England Journal of Medicine, PMID 1454084). Nobody was lying. Human perception of intake is simply bad, and a quiet appetite makes people stop paying attention.

Second, you burn less because there is less of you. A body 15% smaller spends fewer calories at rest and per step. Third, metabolism adapts beyond what the smaller size predicts. The Biggest Loser follow-up found resting metabolic rates roughly 500 kcal/day below expected six years after massive weight loss (Fothergill et al., 2016, Obesity, PMID 27136388). That study involved extreme, rapid loss without medication, so the magnitude likely overstates the typical GLP-1 case, but the direction is universal. Deficits shrink themselves.

Add the quietest leak of all: non-exercise movement. Fidgeting, errands, pacing, stairs. It declines during weight loss without permission, and on appetite-suppressing medication, energy and spontaneous movement often dip together.

The four-week audit to run before blaming the drug

Before any conversation about dose changes, run a four-week audit. Most plateaus do not survive it.

Week one, protein and calories on paper. Weigh and log everything for seven days, including oils, dressings, bites while cooking, and every liquid. Liquid calories, lattes, juices, alcohol, are the most common leak we find. Compare against your actual target, not the target from eight months and 15 kg ago; a smaller body needs a recalculated deficit.

Weeks one to four, a step floor. Pick a number you hit on your worst day, typically 7,000 to 8,000, and defend it daily. This patches the NEAT leak directly.

Weeks one to four, strength training logged. Two to three sessions weekly, loads written down. If lifts are getting weaker while weight is flat, the problem is muscle loss and recovery, not insufficient drug.

Week four, judge the trend, not the day. Use the 7-day rolling average of morning weight, plus waist measurement. Water masks fat loss for weeks at a time, especially around training changes and menstrual cycles.

If the audit comes back clean — calories verified, steps held, training progressing — and the trend is still flat after four to six weeks, then the dose conversation belongs to your prescriber. Titration is their call, never a self-serve menu. But bring the audit data to that appointment; it changes the quality of the decision.

And ask the better question first: flat at what? A plateau at a waist circumference that no longer carries metabolic risk, with strength climbing and bloodwork clean, is not a stall. For many people it is the destination wearing a disguise.

What this looks like in coaching

Plateau consults at the practice follow a strict order: data before dose, always. The client runs the four-week audit while we watch three numbers weekly: intake average, step average, and main-lift loads.

Roughly two-thirds of GLP-1 plateaus we see resolve inside the audit itself, usually from rediscovered liquid calories, portion drift, or a step count that fell off a cliff when the weather turned. The remainder split between true physiological equilibrium, where we discuss maintenance as a win, and cases we refer back to the prescriber with four weeks of clean data in hand.

The scale stalling is information, not verdict. Audit first. Most of the time, the leak has a name and a fix.