When a GLP-1 does not seem to work: non-responders, wrong dose, or wrong expectations.

The short answer

A GLP-1 not working is real but rarer than it feels: in the branded trials, 9.1 percent of adults on the highest tirzepatide dose and 16.5 percent on semaglutide 2.4 mg failed to lose even 5 percent of body weight, which means the large majority responded, and most "it is not working" stories turn out to be a dose, a calendar, a storage problem, or an expectation rather than a non-response. The clinician's job, and this article's, is telling those apart in the right order.

Here is the sorting, from the definitions to the boring checks to the honest conversation about switching.

What counts as a response, by the numbers

Before declaring a medication failed, define what working means, because the trials already did.

The standard clinical benchmark is 5 percent of body weight, the threshold where metabolic health measurably improves. Against it, the branded labels' trial tables read as follows: on semaglutide 2.4 mg, 83.5 percent of adults reached a 5 percent loss over 68 weeks versus 31.1 percent on placebo, with the average participant down 14.9 percent. On tirzepatide over 72 weeks, 85.1 to 90.9 percent reached 5 percent across the maintenance doses versus 34.5 percent on placebo, with averages from 15.0 to 20.9 percent. Flip those responder rates and you get the honest minority: 16.5 and 9.1 percent respectively who did not reach the threshold on the studied doses. Trials of branded, FDA-approved products, not compounded preparations, and populations rather than individuals, but they set the base rate: true limited response happens, roughly one person in ten on the strongest studied regimen, and it is nobody's fault when it does.

Two calibration notes before you place yourself in that minority. The trial figures above are endpoints after a year-plus at maintenance doses, not verdicts on an early month. And "working" includes more than the scale: appetite quieter, portions smaller, the food chatter turned down. If none of that is present either, keep reading; if it is, you are likely mid-story, not at its end.

The first 12 weeks are not the verdict

The most common "GLP-1 not working" case is a calendar misunderstanding, because the early protocol is designed to look underwhelming.

Titration starts at doses the labels themselves describe as initiation rather than treatment: four weeks at 0.25 mg of semaglutide or 2.5 mg of tirzepatide are on-ramps, there to build gut tolerance, and meaningful appetite effect often arrives only in the middle rungs. Add the arithmetic of long half-lives, each new dose takes weeks to reach its full level, and the honest window for judging the medication is after roughly 12 weeks, at a working dose, not during the on-ramp. We wrote the twelve-week stall guide for exactly this checkpoint, and the full results timeline maps the arc beyond it.

Clinically, the 12-week mark at a therapeutic dose is also where the formal definitions of inadequate response begin, not week three at a starter dose. If your disappointment predates a single week at a middle rung, the treatment has not started yet in any pharmacological sense. Patience here is not consolation; it is the protocol.

Expectations deserve their own honest sentence, because the comparison group most people carry is not the trials; it is social media. The branded trials' celebrated averages, 14.9 and 20.9 percent, took 68 and 72 weeks to accumulate, which works out to steady fractions of a percent per week, and half of every trial population, by definition, did worse than the average. A member losing a pound every week or two at month three is not failing against the data; they are the data. What the highlight reels do not show is the timescale, and recalibrating to the trials' actual pace resolves a meaningful share of "not working" before any other check is needed.

Dose, storage, and technique: the boring checks

When response is genuinely absent, clinicians check the plumbing before the pharmacology, because the plumbing fails more often.

Dose: which rung, exactly, and for how many weeks settled? An appetite effect missing at 0.5 mg of semaglutide is expected, not alarming; the question is whether the ladder has actually been climbed per the dose chart, with four settled weeks per rung and no missed-dose gaps quietly resetting adaptation.

Storage: these are fragile proteins. A vial that froze against a fridge's back wall, cooked in a car, or outlived its dates can lose potency with no visible change, and a compounded vial adds its beyond-use date to the checklist. If response vanished with a new vial, or never existed with a vial whose history is fuzzy, the storage rules are suspect number one.

Technique and units: subcutaneous means into the fat layer, sites rotated, the full dose delivered; and on compounded vials, the units-to-milligrams conversion runs through that vial's printed concentration. A syringe drawn to the wrong line delivers the wrong dose with perfect consistency. Your protocol card and your clinician settle this in minutes, and no forum chart should ever be consulted instead.

These checks are unglamorous, and they resolve a meaningful share of "not working" messages without a single prescription change.

One vial-change pattern deserves special mention because it looks exactly like sudden failure: response that was present for months and then vanished within a week or two of starting a new shipment. That timing points hard at the supply chain rather than your physiology, a vial that warmed in transit, a different concentration read with the old vial's math, or a storage slip after arrival, and it is the single fastest "not working" case a clinician can solve. Date the change, keep the vial in question, and send the timeline; the fix is usually a replacement and a recheck, not a new protocol.

Eating around the medication

The medication quiets appetite; it does not stand between you and the refrigerator. There is a recognizable pattern clinicians call eating around the drug, and it deserves a paragraph without a milligram of moralizing.

A GLP-1 mutes hunger signals, but calories that arrive without requiring hunger still count: grazing that never triggers fullness, liquid calories that bypass the slowed stomach's signals entirely, alcohol's quiet arithmetic, and the soft foods that stay easy when appetite is low. Someone can be in honest energy surplus with a perfectly suppressed appetite, and the scale reports the surplus, not the suppression. This is not weakness; it is the gap between an appetite tool and a physics problem, and it responds to structure rather than willpower: protein anchoring each meal, calories that require chewing, and a plate built for the medication.

The tell, for what it is worth: appetite effects clearly present, portions visibly smaller at meals, yet weight flat across six honest weeks. That combination points at the margins of the day, and a food log shared with your clinician, three ordinary days, no performance, finds the leak faster than any dose change would.

True non-response and the switch decision

After the calendar is respected, the plumbing checked, and the plate examined, a residue of genuine limited response remains, the trials' one in ten, and the conversation changes from troubleshooting to strategy.

The first option is the other molecule. Tirzepatide works through two receptor pathways rather than one, and in the head-to-head trial published in the New England Journal of Medicine in 2025 it outperformed semaglutide 2.4 mg on average, 20.2 versus 13.7 percent over 72 weeks. Individual scatter is wide in both arms, and a population average is not a personal forecast, but a limited semaglutide response is among the clearest reasons clinicians consider switching to tirzepatide, started at a rung chosen from your history rather than from zero.

The second option is honesty about goals and adjuncts: for some, the medication's contribution is modest and the protocol leans harder on training, protein, and sleep; for others, a different class or a comprehensive program conversation is the right referral. What a member owes the decision is data, six settled weeks at a real dose, honestly measured; what the clinician owes back is a plan that never mistakes a base-rate reality for a personal failing.

“Before I call anyone a non-responder, I check the dose, the vial, the syringe, and the calendar. Most of the time, one of those four confesses.”

Dr. Marcus Okafor, MD

When to message your clinician

Message through the portal when: you have completed four settled weeks at a middle rung with no change in appetite at all; the scale has not trended across six weeks at a stable working dose; response disappeared abruptly with a new vial or shipment; you are unsure your units match your prescribed milligrams; or you have walked this article and want the switch conversation. Bring the specifics that make the first reply useful: current dose and weeks at it, vial concentration if compounded, a few days of honest eating, and how appetite actually feels at 4 pm. At Parke a licensed clinician reads every one of these and answers with a plan, because a protocol that is not working is precisely what the 24-hour review exists to catch early.

The bottom line

A GLP-1 not working is a real phenomenon with a measured size: 9.1 percent short of the 5 percent threshold on the strongest branded tirzepatide dose, 16.5 percent on branded semaglutide 2.4 mg, figures from the FDA-approved products' trials, not compounded preparations. Everything outside that minority is findable: a dose still on its on-ramp, a calendar judged too early, a vial that lost the cold chain, a syringe reading someone else's math, or calories arriving around the appetite rather than through it. The sequence is patience to twelve weeks at a working dose, the boring checks, the plate, and only then the switch.

If you want that sequence run by someone accountable for the answer, tirzepatide and semaglutide at Parke both come with the clinician attached, and eligibility takes about two minutes.

GLP-1 not working: frequently asked questions

How long before I know if semaglutide is working?

Judge at about 12 weeks, including several settled weeks at a working middle dose, not during the 0.25 mg on-ramp, whose job is tolerance rather than results. Appetite changes often arrive before scale changes. If twelve honest weeks at a therapeutic rung show neither, that is a real signal, and exactly the message your clinician wants to receive.

What percentage of people do not respond to GLP-1s?

Using the trials' 5 percent threshold, 16.5 percent of adults on branded semaglutide 2.4 mg and 9.1 percent on the highest branded tirzepatide dose did not reach it, versus roughly two thirds of placebo participants. So genuine limited response is real but uncommon, about one in ten on the strongest studied regimen, and most apparent failures trace to dose, timing, storage, or intake instead.

Should I switch to tirzepatide if semaglutide is not working?

Often yes, after the checks. If a true limited response survives twelve weeks at a working dose with storage, technique, and eating examined, tirzepatide's dual-pathway mechanism and stronger head-to-head averages make it the natural next move. The switch is a clinician decision, started at a rung chosen from your history; it is routine at Parke and worth a message.

This article is general education, not medical advice, and it is not a complete list of risks or side effects. Trial figures cited here come from studies of branded, FDA-approved products; compounded medications are prepared by state-licensed US compounding pharmacies and are not FDA-approved or evaluated for safety, effectiveness, or quality. Whether any treatment is appropriate for you is a decision made with a licensed clinician who knows your history. If you are experiencing a medical emergency, call 911. Review the important safety information on each treatment page.