The GLP-1 Non-Responder Problem: Why These Drugs Barely Work for Some People

GLP-1 drugs are usually described through averages. The average patient loses a certain percentage of body weight. The average appetite becomes quieter. The average clinical-trial participant reaches a particular result after 68 or 72 weeks. But no real person is an average. One patient begins semaglutide and quickly notices that the constant mental conversation about food has almost disappeared. Another follows treatment for months, deals with nausea and rising costs, and watches the scale barely move. Both experiences are real. These medications can be extraordinarily effective. They are not equally effective for everyone. And the reason is more complicated than "the patient did something wrong." The Numbers Behind the Non-Responder Problem The table below gives a clearer picture than headline averages alone. The semaglutide figures come from Wegovy clinical studies, while the tirzepatide figures come from separate Zepbound studies. These were not head-to-head comparisons. Their populations, protocols, escalation schedules, treatment durations, and statistical methods differed. Tirzepatide is shown across all three maintenance doses studied for weight management: 5 mg, 10 mg, and 15 mg. The frequently quoted 20.9% result occurred with 15 mg, the maximum recommended weekly dose. The classic semaglutide results shown here were obtained using 2.4 mg weekly in the pivotal studies. The current U.S. Wegovy label also includes a 7.2 mg injection option for certain adults with obesity, but that newer regimen is not represented by the 14.9% figure in this table. Milligram values between the two drugs are not measures of equivalent pharmacological strength. Fifteen milligrams of tirzepatide should not be interpreted as directly comparable with 2.4 milligrams of semaglutide simply because the numbers appear in the same table. Still, the table reveals the central issue: even under highly controlled trial conditions, a meaningful minority of participants did not reach 5% weight loss. What Does "Non-Responder" Actually Mean? There is no single perfect definition. In obesity research and clinical practice, losing at least 5% of starting body weight is commonly treated as a clinically meaningful milestone. A person who remains below that threshold may be described as a low responder or non-responder. But timing changes everything. Losing less than 5% during the first weeks of treatment is not the same as losing less than 5% after a sustained period at the highest appropriate and tolerated maintenance dose. GLP-1 treatment usually begins with a low introductory dose. The dose is then increased gradually to reduce gastrointestinal adverse effects. Tirzepatide trials included escalation periods lasting as long as 20 weeks, while semaglutide injection is also increased in stages rather than started at its maintenance dose. A person may therefore say, "I have been taking it for four months," even though much of that time was spent moving through introductory doses. Calendar time is not the same as adequate treatment exposure. At the opposite extreme, waiting indefinitely for a dramatic transformation is not a useful strategy either. A fair evaluation asks three questions: Has the person had enough time? Have they reached a clinically appropriate and tolerable dose? Has the treatment been consistent enough to judge? Without that context, the label "non-responder" can be premature. Why Apparent Treatment Failure Is Often More Complicated Than It Looks Several separate issues can make an effective medication appear ineffective. Dose, duration, and interruptions can distort the picture Treatment may be interrupted by shortages, insurance denials, unaffordable refills, side effects, missed appointments, or difficulty obtaining the same product consistently. A person who repeatedly starts, stops, restarts, or changes formulations has not had the same exposure as a participant in a tightly controlled clinical trial. That does not mean the patient has failed. It means the treatment environment has been unstable. "GLP-1" is a category, not one interchangeable drug Semaglutide activates the GLP-1 receptor. Tirzepatide activates both GIP and GLP-1 receptors. Older medications such as liraglutide have different dosing schedules and pharmacological profiles. Oral and injectable semaglutide products also differ in administration, absorption, approved uses, and dose ranges. Even medications that influence related metabolic pathways are not interchangeable. In the head-to-head SURMOUNT-5 trial among adults with obesity or overweight without type 2 diabetes, tirzepatide produced greater average weight reduction than semaglutide. That does not mean tirzepatide is automatically the correct medication for every individual. The strongest drug on a clinical-trial graph is not useful to someone who cannot tolerate it, cannot access it consistently, or has a medical reason not to use it. A plateau is not automatically a failure Weight loss rarely continues at the same speed forever. As body weight falls, energy requirements change. Early fluid shifts stop contributing to scale movement. Appetite and energy expenditure adapt. Eventually, many patients reach a new, relatively stable weight. A person who loses 15% and then maintains that result has not suddenly become a non-responder. The medication may be helping maintain the new weight rather than producing continued weekly losses. Obesity medications generally manage weight-regulation biology while they are being used; discontinuation is commonly followed by regain rather than permanent preservation of the drug-induced result. The more relevant question is not simply, "Why did the scale stop?" It is: Dose, duration, drug selection, interruptions, and plateaus can all alter how treatment appears. They should be separated from true biological non-response. The Drug Changes Appetite. It Does Not Erase Calories. This is the uncomfortable section that GLP-1 discussions often avoid. Semaglutide can reduce appetite and calorie intake. Tirzepatide can make smaller portions feel satisfying. Food noise may become quieter, and the urge to snack may weaken. But the medications do not make energy intake irrelevant. Someone may eat less frequently while still consuming enough energy to prevent substantial weight loss. Small quantities of highly energy-dense foods can contain more calories than larger meals built around less calorie-dense foods. Sugary drinks, specialty coffees, alcohol, sauces, oils, desserts, and repeated small snacks may contribute considerable energy without feeling like traditional meals. This creates a confusing experience: The statement may feel completely true. The person may genuinely be eating fewer meals and experiencing much less hunger. But reduced meal frequency does not always produce a meaningful reduction in total energy intake. This is not a moral judgment. Obesity is not caused by a lack of character, and GLP-1 treatment should not become another reason to blame patients. Appetite biology, food environment, sleep, stress, mobility, other medications, economic limitations, and long-established habits all interact. Still, the role of behavior cannot be edited out of the science. Wegovy and Zepbound are indicated in combination with a reduced-calorie diet and increased physical activity. In the pivotal tirzepatide program, participants also received structured lifestyle counseling rather than medication in isolation. The practical review is not, "Did the patient try hard enough?" It is more specific: Has appetite actually decreased? Have portion sizes changed? Are liquid calories still present? Has alcohol intake changed? Is the person eating less food by volume but choosing more energy-dense food? Is persistent nausea leading to irregular daytime eating followed by evening compensation? Is poor sleep increasing hunger and reducing activity? A medication can create an opportunity for behavioral change. It cannot decide how that opportunity is used. Do not overlook declining daily movement Food intake is only one side of the energy-balance equation. Non-exercise activity thermogenesis, usually shortened to NEAT, includes the energy used for ordinary movement outside formal exercise: walking around the house, standing, climbing stairs, doing chores, changing posture, gesturing, and even fidgeting. Fatigue is reported among the common adverse reactions associated with both semaglutide and tirzepatide weight-management treatment. When fatigue occurs, daily movement may fall without the person consciously deciding to become less active. Someone may walk fewer steps, spend more time sitting, stop making short journeys on foot, or perform fewer household tasks. That decline can quietly offset part of the energy deficit created by appetite suppression. A person may genuinely be eating less while also expending less energy than before. That does not mean thermodynamics has stopped working or that the patient is secretly overeating. It means both sides of the equation have changed. The exact reduction cannot be assumed to be 200 or 300 calories for every GLP-1 user. NEAT varies enormously between people, and a drug-specific average of that kind has not been established. Research does show, however, that caloric restriction and weight loss can reduce both resting and activity-related energy expenditure. A simple practical check is to compare average daily steps and sedentary time across several weeks before and during treatment. The goal is not to treat a smartwatch calorie estimate as a laboratory measurement. It is to notice whether ordinary movement has fallen substantially. The more useful review therefore includes both questions: Has total energy intake actually decreased? Has fatigue or weight-loss adaptation also caused daily activity to decrease? Protecting regular walking, standing, household movement, and appropriately prescribed resistance exercise may help preserve physical function and prevent appetite suppression from being partially cancelled by an increasingly sedentary day. Tolerability Can Cap the Dose That Is Actually Usable A medication cannot produce its theoretical trial result when the patient cannot tolerate enough of it to remain on treatment. Nausea, vomiting, diarrhea, constipation, reflux, abdominal discomfort, and fatigue are common with GLP-1–based therapies. For some people, the symptoms improve after dose escalation. For others, they remain disruptive enough to delay increases, force dose reductions, or end treatment. In pooled Zepbound weight-reduction studies, gastrointestinal reactions were reported in 56% of tirzepatide-treated participants compared with 30% receiving placebo. Most nausea, vomiting, and diarrhea events occurred during escalation and decreased over time, but some participants discontinued because of gastrointestinal effects. This can create a particularly frustrating middle ground. The person receives enough medication to feel nauseated, tired, constipated, or uninterested in food—but not enough, or not consistently enough, to experience the average weight loss seen in trials. That is not necessarily receptor-level resistance. It may be a tolerability ceiling. It is also why severe side effects should not be interpreted as proof that the drug is "working harder." More vomiting does not mean more fat loss. Persistent gastrointestinal symptoms can cause dehydration, kidney complications, nutritional disruption, and treatment discontinuation. The most effective dose is not automatically the highest available dose. It is the dose that produces meaningful benefit while remaining medically safe and realistically tolerable. Type 2 Diabetes Changes the Expected Response Semaglutide and tirzepatide were not originally developed as celebrity weight-loss injections. They emerged from diabetes treatment. Paradoxically, people with type 2 diabetes tend to lose less average weight on these medications than people without diabetes. In the Wegovy clinical data, participants without diabetes lost an average of 14.9% of their starting body weight, while participants with type 2 diabetes lost an average of 9.6%. The proportion reaching at least 5% weight loss fell from 83.5% to 67.4%. The same pattern appears in tirzepatide trials. Participants without diabetes receiving 15 mg lost an average of 20.9%, while those with type 2 diabetes receiving 15 mg lost an average of 14.7% in a separate trial. The 2026 Nature analysis also found type 2 diabetes to be a strong predictor of lower weight-loss efficacy, with an average predicted reduction of 2.87 percentage points in BMI loss after adjustment for other measured factors. Researchers are still examining the full explanation. Diabetes duration, insulin resistance, pancreatic function, glucose-lowering medications, metabolic adaptation, and differences between trial populations may all contribute. The practical implication is simple: Different metabolic contexts produce different averages. That does not mean the medication has failed. Weight loss may be more modest while glucose control or other clinically important outcomes improve. Genetics Matters—but It Is Not Yet a Dosing Manual This is the most scientifically interesting part of the non-responder problem. A 2026 Nature study examined self-reported treatment response and side effects among 27,885 people who had used semaglutide, tirzepatide, or related products. Researchers identified a variant in the GLP1R gene associated with weight-loss response and found associations involving both GLP1R and GIPR with nausea or vomiting. The GIPR association was specific to people using tirzepatide. One GLP1R variant was associated with approximately 0.76 kilograms of additional weight loss per copy of the effect allele. That is scientifically meaningful. It is not a genetic verdict. The difference is far too small to fully explain why one person loses 25% of their body weight while another loses almost nothing. When the researchers combined genetics with demographic and clinical variables, medication details, dosage, and treatment duration, their model explained approximately 25% of the variation in self-reported BMI loss. Most of that predictive ability came from non-genetic factors. The study also had important limitations. The information was largely self-reported. Approximately 82% of respondents were women, and 78% were of European ancestry. Recorded electronic health data showed smaller weight changes than participants reported in surveys. So what can genetics do right now? It can offer a clue. The findings suggest that future prescribing models may help clinicians estimate which medication is more likely to work, who may be more vulnerable to nausea or vomiting, and who may benefit from a more cautious escalation strategy. But the study does not establish a validated genotype-based dosing protocol. A consumer DNA result should not currently be used to independently select a starting dose, accelerate escalation, or override approved prescribing instructions. In current practice, titration decisions remain driven primarily by medical history, actual side effects, treatment response, and tolerability. The practical bridge is this: It may eventually help clinicians choose between drugs and escalation strategies before patients spend months discovering the answer through trial and error. The research represents a foundation for precision obesity medicine—not a finished clinical tool. The Scale May Not Capture the Entire Response Weight matters, but it is not the only outcome worth measuring. A person taking a GLP-1 medication for type 2 diabetes may experience meaningful improvement in blood glucose despite modest weight loss. Another person may notice:

• reduced hunger or food noise;
• fewer episodes of uncontrolled eating;
• a smaller waist circumference;
• improved mobility;
• changes in blood pressure;
• better sleep or metabolic markers.

These changes do not mean that every weak scale response justifies continuing an expensive medication indefinitely. They mean success should be evaluated against the original treatment goal. Was the medication prescribed primarily for chronic weight management, diabetes control, cardiovascular-risk reduction, obstructive sleep apnea, or another approved purpose? A person can be a low weight-loss responder while still receiving another clinically meaningful benefit. That distinction matters when deciding whether treatment should continue. When Poor Response Becomes a Reason to Reassess A slow start is not automatically a failure. But a medication should not be continued forever without evidence that it is helping. There is no universal FDA rule stating that less than 1% weight loss after six months means treatment must be stopped. That figure should not be presented as a recognized clinical standard. Different guidelines use different review points. A longstanding Endocrine Society guideline recommends assessing obesity-medication efficacy and safety at least monthly for the first three months. It describes weight loss of at least 5% after three months as an effective response and recommends reconsidering treatment when the response is below that level or when safety or tolerability problems occur. That guideline predates modern, extended GLP-1 and GIP/GLP-1 escalation schedules, so the three-month benchmark must be interpreted carefully. A patient who is still in low-dose titration has not had the same exposure as someone who has spent three months at a stable therapeutic dose. NICE guidance provides a later, drug-specific benchmark for tirzepatide: when less than 5% of initial weight has been lost after six months on the highest tolerated dose, the patient and clinician should decide whether continuing treatment remains worthwhile after considering its benefits and risks. These are review thresholds—not permission to change treatment alone. A careful reassessment should examine:

• whether the medication was used consistently;
• how long the person spent at a therapeutic or highest tolerated dose;
• whether side effects prevented adequate exposure;
• whether total energy intake changed;
• whether fatigue or reduced NEAT lowered daily expenditure;
• whether another medication or health condition is affecting weight;
• whether benefits occurred beyond the scale;
• whether the cost and burden remain justified.

When meaningful benefit is absent after an adequate and consistent trial, continuing the same strategy simply because "maybe it will start working next month" may waste time and money. That is the moment for a structured treatment review—not an improvised dose increase. Red Flags: When This Is No Longer Just a Weight-Loss Question Poor efficacy and medical danger are not the same thing. Losing less weight than expected usually calls for a scheduled review. Certain symptoms require prompt or urgent medical attention. Persistent or severe abdominal pain Severe abdominal pain that does not go away—particularly pain that radiates toward the back and occurs with or without vomiting—can be a sign of pancreatitis. Current Zepbound prescribing information instructs patients to discontinue the medication when pancreatitis is suspected and obtain appropriate medical management. Upper abdominal pain with jaundice, fever, or pale stools Yellowing of the skin or eyes is not simply a sign that the medication is "too strong." Together with upper abdominal pain, fever, or pale or clay-colored stools, it may indicate a gallbladder or bile-duct problem and requires medical evaluation. Gallbladder disease, including cholelithiasis and cholecystitis, is included among the warnings for these treatments. Vomiting or diarrhea that does not stop Persistent vomiting or diarrhea can lead to dehydration and kidney injury. Inability to keep fluids down, severe weakness, dizziness, fainting, or substantially reduced urination should not be managed by simply delaying or changing the next dose without professional advice. Signs of a serious allergic reaction Swelling of the face, lips, tongue, or throat; difficulty breathing or swallowing; fainting; or a severe rash requires immediate medical help. Serious hypersensitivity reactions, including anaphylaxis and angioedema, have been reported with tirzepatide and other GLP-1–based products. These are not normal signs that the drug is working. They are reasons to stop experimenting and seek medical assessment. Can Switching From Semaglutide to Tirzepatide Help? Possibly. Because semaglutide and tirzepatide do not have identical mechanisms, a weak response to one does not prove that every incretin-based medication will fail. Head-to-head evidence shows that tirzepatide produces greater average weight loss than semaglutide among adults with obesity or overweight without type 2 diabetes. But switching is not the same as upgrading a smartphone. It may introduce:

• different gastrointestinal side effects;
• a new escalation period;
• insurance or supply barriers;
• different contraindication considerations;
• another period of uncertain individual response.

The new medication should also receive a fair trial. Tirzepatide normally begins at a low initiation dose and is increased gradually. The first weeks are designed largely to establish tolerability, not to demonstrate the maximum possible weight-loss effect. Its pivotal weight-management studies included escalation periods of up to 20 weeks. A switch should therefore not be declared a failure simply because the scale has barely moved during the introductory phase. A review at approximately 12 weeks can be useful as an early checkpoint for adherence, appetite response, glucose changes when relevant, side effects, and the overall direction of treatment. A prospective study followed people with type 2 diabetes for 12 weeks after they switched directly from another stable GLP-1 receptor agonist regimen to tirzepatide 5 mg. Participants experienced additional average weight reduction and improved glycemic outcomes, although gastrointestinal events occurred in 13.2%, and the study was not designed to establish a universal 12-week stopping rule for obesity treatment. Twelve weeks is therefore a useful review point—not a deadline for the final result. It is particularly important not to judge the switch only by time spent at low introductory doses. A practical timeline looks more like this: During titration: evaluate safety, side effects, adherence, and whether appetite is beginning to change. After reaching a stable maintenance dose: assess the direction and pace of the response rather than expecting an immediate final result. At the formal efficacy review: decide whether the benefit is large enough to justify continued cost, adverse effects, and treatment burden. For tirzepatide, NICE recommends reconsidering continuation when less than 5% of initial weight has been lost after six months on the highest tolerated dose. A person who did not respond well to semaglutide may respond better to tirzepatide. Another may experience severe side effects and obtain no additional benefit. The decision should be based on the reason for the poor response. Was the problem biological non-response? Was the dose limited by nausea? Was treatment inconsistent? Did daily movement fall because of fatigue? Did the person remain hungry, or did calorie intake stay high despite reduced hunger? Was the medication being used primarily for diabetes rather than weight management? Switching drugs without answering those questions may simply reproduce the same disappointment under a different brand name. The Future Is Precision Obesity Medicine For years, obesity pharmacotherapy has relied heavily on trial and error. Start a medication. Increase the dose. Wait. Measure the response. Manage the side effects. Continue, switch, or stop. That approach is beginning to change. Researchers are now combining genetic variants with medication type, dosage, treatment duration, age, sex, disease history, and early response patterns. Future tools may help predict:

• who is most likely to respond to semaglutide;
• who may do better with tirzepatide or another metabolic pathway;
• who is at higher risk of severe gastrointestinal effects;
• who may benefit from slower escalation;
• who is unlikely to receive enough benefit to justify the cost.

But current models still leave most of the individual outcome unexplained. The 2026 Nature model explained approximately one-quarter of the observed variation in self-reported BMI loss, meaning the majority remained outside the model's predictive reach. The science can estimate what may happen to a population. It still cannot promise what will happen to one person. The Bottom Line GLP-1 medications work. That statement is supported by large clinical trials and by life-changing outcomes for many patients. But "works" is not a binary switch. Some people lose more than 20% of their body weight. Some lose 5% to 10%. Some improve their blood glucose while losing relatively little weight. Some cannot tolerate the medication. Some receive an unstable or inadequate treatment trial. Some experience enough fatigue or metabolic adaptation that daily energy expenditure falls. Some continue consuming enough energy to offset much of the appetite effect. And some appear to receive very little meaningful benefit despite giving the treatment a fair chance. The GLP-1 non-responder problem does not prove that these drugs are fraudulent, useless, or universally overhyped. It proves something less dramatic and more important: The next phase of obesity medicine will not be defined only by stronger drugs. It will be defined by learning how to identify the right drug, the right treatment environment, and the right expectations for the individual patient—before months of side effects, cost, and disappointment are required to discover the answer.