Obesity looks like one diagnosis on paper, but it doesn't behave like one disease. Some people can eat a very large meal before feeling full. Others feel full on a normal portion but are hungry again an hour later. Some eat in response to stress or emotion, and some burn fewer calories at rest than expected. Researchers at Mayo Clinic have formalized these patterns into four obesity phenotypes, and early evidence suggests that which pattern you have may help explain why the same GLP-1 produces wildly different results in different people.
This guide covers the four phenotypes, what the studies show about GLP-1 response, and what to know before spending money on a phenotype test.
The four phenotypes
In a study of 450 adults with obesity, Mayo Clinic researchers measured satiation, satiety, eating behavior, mood, and resting metabolism, and found four recurring patterns (Acosta et al., Obesity, 2021, DOI (external link)):
- Hungry brain (abnormal satiation). The "I'm full" signal arrives late, so you can eat well past a normal meal's worth of calories before feeling satisfied.
- Hungry gut (abnormal satiety). You feel full on a normal meal, but the fullness doesn't last. The stomach empties faster than usual and hunger returns quickly, often driving frequent snacking.
- Emotional hunger (hedonic eating). Eating is driven by emotion, stress, or reward instead of physical hunger.
- Slow burn (decreased metabolic rate). Resting energy expenditure is lower than predicted, so weight accumulates even at ordinary intakes.
About 85% of participants fit at least one phenotype, roughly a quarter fit more than one, and 15% fit none, so the map is useful but incomplete. Having two or more phenotypes stacks: in a later cohort of 464 patients, people with multiple phenotypes carried significantly higher weight and BMI than people with one or none (Ghusn et al., International Journal of Obesity, 2024, DOI (external link)). The biology also appears to run in the direction you'd expect. In younger adults followed for about four years, those with faster stomach emptying at baseline gained roughly three times as much weight (Pajot et al., International Journal of Obesity, 2020, DOI (external link)).
Why phenotype could plausibly matter for GLP-1s
GLP-1 medications work mostly on the systems the "hungry gut" phenotype describes: they slow stomach emptying and amplify the gut-to-brain fullness signal. If your obesity is driven by a stomach that empties too fast and fullness that fades too soon, a drug that corrects both is aimed at your actual problem. If it's driven mainly by emotional eating or a low metabolic rate, the same drug is aimed somewhere else. It may still help, but there's less reason to expect a dramatic response.
So far, the evidence mostly supports that idea.
What the studies show
Phenotype-guided prescribing beat trial-and-error. In a 12-month pragmatic trial of 312 patients, those whose anti-obesity medication was chosen to match their phenotype (for example, a GLP-1 for hungry gut) lost 15.9% of body weight versus 9.0% with standard non-phenotype-guided selection. That's a 1.75-fold difference, with 79% versus 34% losing more than 10% (Acosta et al., Obesity, 2021, DOI (external link)).
A "hungry gut" subtype nearly doubled tirzepatide results. In 2026, the same group profiled 483 adults and found that about 1 in 4 had a distinctive subtype: fast stomach emptying, more post-meal hunger, and, surprisingly, low natural GLP-1 and PYY levels, traced to reduced hormone production in the gut lining itself. In a retrospective subset treated with tirzepatide, this group lost 21.5% of body weight at six months versus 11.7% for everyone else (Ticho et al., Gastroenterology, 2026, DOI (external link)). The people whose bodies make the least GLP-1 may benefit the most from replacing it.
The idea extends beyond medication. A 12-week proof-of-concept trial found phenotype-tailored lifestyle programs produced −7.4 kg versus −4.3 kg for a standard program (Cifuentes et al., eClinicalMedicine, 2023, DOI (external link)), and reviews from the same program lay out the broader precision-medicine case (Anazco & Acosta, International Journal of Obesity, 2024, DOI (external link)).
The important caveats
Before treating any of this as settled, know three things:
- Almost all of the GLP-1-specific outcome data comes from one research group. The phenotype framework, the guided-prescribing trial, and the tirzepatide subtype study all come from Mayo Clinic's Precision Medicine for Obesity Program. That doesn't make the work wrong (it's peer-reviewed and published in strong journals), but no one has independently replicated the outcome data yet. The taxonomy itself has spread beyond Mayo: a 2025 international review from groups in Italy, Romania, and the UAE surveys the same four phenotypes as an established framework for obesity management (Patti et al., Medicina, 2025, DOI (external link)). That independent review is also explicit that no standardized, phenotype-specific treatment protocols exist yet, including for diet. It calls phenotype-tailored nutrition "hypothetical," a proposed approach that hasn't been tested.
- The headline GLP-1 numbers are retrospective and small. The 21.5%-vs-11.7% tirzepatide finding came from looking back at 61 treated patients, not from a randomized trial. The guided-prescribing trial used liraglutide, an older, weaker GLP-1, alongside four non-GLP-1 drugs.
- There's a commercial product attached. The phenotype framework has been licensed to a company, Phenomix Sciences, which sells a genetic phenotype test. The company reports its test predicts response across treatments, but much of that data comes from conference presentations and company-affiliated studies, not independent evaluation. A financial stake doesn't invalidate the science. It does mean marketing claims deserve extra scrutiny.
The test that exists today, and what it covers
The one commercial implementation of this research is MyPhenome, a cheek-swab genetic test from Phenomix Sciences, the company that licensed Mayo Clinic's phenotype framework. The research studies measured physiology directly (calorie-to-fullness meals, gastric-emptying scans, metabolic-rate testing). MyPhenome instead uses gene-based risk scores to estimate which phenotype your biology leans toward. It has been offered through participating weight-management clinics for a few years, and since June 2026 it can also be ordered directly from the company's website (myphenometest.com (external link)) (Phenomix announcement, June 16, 2026 (external link)). You complete a questionnaire, pay, swab at home, and get results in a secure portal a few weeks later, supported by the company's physician network. It's a one-time out-of-pocket payment (the company lists it as HSA/FSA-eligible). Check the company's site for current pricing and ask your insurer about coverage before assuming any.
Two things to know before ordering it:
- It currently identifies only some of the phenotypes. The consumer test screens for hungry gut, hungry brain, and emotional hunger, but not slow burn, which the company describes as an "emerging" phenotype it plans to add later. In the research setting, slow burn (low resting metabolic rate) is measured with metabolic equipment, not genetics. So a "no phenotype detected" result doesn't rule out a metabolic driver, and it doesn't mean your obesity has no biological basis. The research studies themselves couldn't classify 15% of participants.
- A result is a probability, not a diagnosis. The gene scores estimate which pattern you likely have. They are not the direct physiologic measurements the published trials were built on, and the test's predictive claims haven't yet been independently validated outside company-affiliated studies.
Do you need a phenotype test before starting a GLP-1?
No. Phenotype testing is not part of any standard treatment guideline, and no result, in either direction, proves a GLP-1 will or won't work for you. The strongest predictor available to you costs nothing: your own early response. Weight change in the first three to four months at an adequate dose predicts the eventual outcome reasonably well. In pooled liraglutide 3.0 mg trials, losing at least 4% by week 16 was the best early predictor of reaching 5% or more at one year (Fujioka et al., Obesity, 2016, DOI (external link)). That's why prescribers use an early checkpoint to continue, adjust, or switch. The whole arc is laid out in what to expect on a GLP-1.
For now, the phenotype lens is most useful as a way to frame a conversation. If your pattern sounds like hungry gut (full on normal meals, hungry again fast, frequent snacking), the research suggests you're in the group most likely to respond well to GLP-1s. If your pattern sounds like emotional hunger or hungry brain and a GLP-1 has underdelivered, raise that with your prescriber. Other medications and structured behavioral support target those drivers more directly, and a weak GLP-1 response may simply mean the drug and your main driver don't match. Where the test may earn its cost is the middle ground: you've had a disappointing response at a full dose, the usual fixes haven't explained it, and you and your prescriber want another data point before choosing the next medication. Understanding your own hunger pattern is free, and it can make that conversation sharper.
This article is general education, not medical advice. Phenotype findings are group-level research results and don't predict any individual's outcome; decisions about testing, starting, switching, or stopping medication belong with your prescriber. glp1.how has no relationship with any test manufacturer. Research findings above are attributed to PubMed-indexed articles with DOI links.