Emerging research indicates that phenotype-based testing may help identify which biologic process is driving an individual’s obesity, enabling clinicians to better tailor antiobesity medication (AOM) to the patient.
Currently, patient response to AOMs varies widely, with some patients responding robustly to AOMs and others responding weakly or not at all.
For example, trials of the GLP-1 semaglutide found that 32%-39.6% of people are “super responders,” achieving weight loss in excess of 20%, and a subgroup of 10.2%-16.7% of individuals are nonresponders. Similar variability was found with other AOMs, including the GLP-1 liraglutide and tirzepatide, a dual GLP-1/glucose-dependent insulinotropic polypeptide receptor agonist.
Studies of semaglutide suggest that people with obesity and type 2 diabetes (T2D) lose less weight on the drug than those without T2D, and men tend to lose less weight than women.
However, little else is known about predictors of response rates for various AOMs, and medication selection is typically based on patient or physician preference, comorbidities, medication interactions, and insurance coverage.
Although definitions of a “nonresponder” vary, the Endocrine Society’s latest guideline, which many clinicians follow, states that an AOM is considered effective if patients lose more than 5% of their body weight within 3 months.
Can nonresponders and lower responders be identified and helped? Yes, but it’s complicated.
“Treating obesity effectively means recognizing that not all patients respond the same way to the same treatment, and that’s not a failure; it’s a signal,” said Andres Acosta, MD, PhD, an obesity expert at Mayo Clinic, Rochester, Minnesota, and a cofounder of Phenomix Sciences, a biotech company in Menlo Park, California.
“Obesity is not a single disease. It’s a complex, multifactorial condition driven by diverse biological pathways,” he told Medscape Medical News. “Semaglutide and other GLP-1s primarily act by reducing appetite and slowing gastric emptying, but not all patients have obesity that is primarily driven by appetite dysregulation.”
Phenotype-Based Profiling
Figuring out what drives an individual’s obesity is where a phenotype-based profiling test could possibly help.
Acosta and colleagues previously used a variety of validated studies and questionnaires to identify four phenotypes that represent distinct biologic drivers of obesity: hungry brain (abnormal satiation), emotional hunger (hedonic eating), hungry gut (abnormal satiety), and slow burn (decreased metabolic rate). In their pragmatic clinical trial, phenotype-guided AOM selection was associated with 1.75-fold greater weight loss after 12 months than the standard approach to drug selection, with mean weight loss of 15.9% and 9%, respectively.
“If a patient’s obesity isn’t primarily rooted in the mechanisms targeted by a particular drug, their response will naturally be limited,” Acosta said. “It’s not that they’re failing the medication; the medication simply isn’t the right match for their biology.”
For their new study, published online in Cell Metabolism, Acosta and colleagues built on their previous research by analyzing the genetic and nongenetic factors that influenced calories needed to reach satiation (Calories to Satiation [CTS]) in adults with obesity. They then used machine learning techniques to develop a CTS gene risk score (CTS-GRS) that could be measured by a DNA saliva test.
The study included 717 adults with obesity (mean age, 41; 75% women) with marked variability in satiation, ranging from 140 to 2166 kcals to reach satiation.
CTS was assessed through an ad libitum meal, combined with physiological and behavioral evaluations, including calorimetry, imaging, blood sampling, and gastric emptying tests. The largest contributors to CTS variability were sex and genetic factors, while other anthropometric measurements played lesser roles.
Various analyses and assessments of participants’ CTS-GRS scores showed that individuals with a high CTS-GRS, or hungry brain phenotype, experienced significantly greater weight loss when treated with phentermine/topiramate than those with a low CTS-GRS, or hungry gut, phenotype. After 52 weeks of treatment, individuals with the hungry brain phenotype lost an average of 17.4% of their body weight compared with 11.2% in those with the hungry gut phenotype.
An analysis of a separate 16-week study showed that patients with the hungry gut phenotype responded better to the GLP-1 liraglutide, losing 6.4% total body weight, compared to 3.3% for those with the hungry brain phenotype.
Overall, the CTS-GRS test predicted drug response with up to 84% accuracy (area under the curve, 0.76 in men and 0.84 in women). The authors acknowledged that these results need to be replicated prospectively and in more diverse populations to validate the test’s predictive ability.
“This kind of phenotype-based profiling allows us to predict which patients are more likely to respond and who might need a different intervention,” Acosta said. “It’s a critical step toward eliminating trial-and-error in obesity treatment.”
The test (MyPhenome test) is used at more than 80 healthcare clinics in the United States, according to Phenomix Sciences, which manufactures it. A company spokesperson said the test does not require FDA approval because it is used to predict obesity phenotypes to help inform treatment, but not to identify specific medications or other interventions. “If it were to do the latter,” the spokesperson said, “it would be considered a ‘companion diagnostic’ and subject to the FDA clearance process.”
What to Do if an AOM Isn’t Working?
It’s one thing to predict whether an individual might do better on one drug vs another, but what should clinicians do meanwhile to optimize weight loss for their patients who may be struggling on a particular drug?
“Efforts to predict the response to GLP-1 therapy have been a hot topic,” noted Sriram Machineni, MD, associate professor at Montefiore Medical Center, Bronx, New York, and founding director of the Fleischer Institute Medical Weight Center at Montefiore Einstein. Although the current study showed that genetic testing could predict responders, such as Acosta, he agreed that the results need to be replicated in a prospective manner.
“In the absence of a validated tool for predicting response to specific medications, we use a prioritization process for trialing medications,” Machineni told Medscape Medical News. “The prioritization is based on the suitability of the side-effect profile to the specific patient, including contraindications; benefits independent of weight loss, such as cardiovascular protection for semaglutide; average efficacy; and financial accessibility for patients.”
Predicting responders isn’t straightforward, said Robert Kushner, MD, professor of medicine and medical education at the Feinberg School of Medicine at Northwestern University and medical director of the Wellness Institute at Northwestern Memorial Hospital in Chicago.
“Despite looking at baseline demographic data such as race, ethnicity, age, weight, and BMI, we are unable to predict who will lose more or less weight,” he told Medscape Medical News. The one exception is that women generally lose more weight than men. “However, even among females, we cannot discern which females will lose more weight than other females,” he said.
If an individual is not showing sufficient weight loss on a particular medication, “we first explore potential reasons that can be addressed, such as the patient is not taking the medication or is skipping doses,” Kushner said. If need be, they discuss changing to a different drug to improve compliance. He also stresses the importance of making lifestyle changes in diet and physical activity for patients taking AOMs.
Often patients who do not lose at least 5% of their weight within 3 months are not likely to respond well to that medication even if they remain on it. “So, early response rates determine longer-term success,” Kushner said.
Acosta said that if a patient isn’t responding to one class of medication, he pivots to a treatment better aligned with their phenotype. “That could mean switching from a GLP-1 to a medication like [naltrexone/bupropion] or trying a new method altogether,” he said. “The key is that the treatment decision is rooted in the patient’s biology, not just a reaction to short-term results. We also emphasize the importance of long-term follow-up and support.”
The goal isn’t just weight loss but also improved health and quality of life, Acosta said. “Whether through medication, surgery, or behavior change, what matters most is tailoring the care plan to each individual’s unique biology and needs.”
The new study received support from the Mayo Clinic Clinical Research Trials Unit, Vivus Inc., and Phenomix Sciences. Acosta is supported by a National Institutes of Health grant.
Acosta is a co-founder and inventor of intellectual property licensed to Phenomix Sciences Inc.; has served as a consultant for Rhythm Pharmaceuticals, Gila Therapeutics, Amgen, General Mills, Boehringer Ingelheim, Currax Pharmaceuticals, Nestlé, Bausch Health, and Rare Diseases; and has received research support or had contracts with Vivus Inc., Satiogen Pharmaceuticals, Boehringer Ingelheim, and Rhythm Pharmaceuticals. Machineni has been involved in semaglutide and tirzepatide clinical trials and has been a consultant to Novo Nordisk, Eli Lilly and Company, and Rhythm Pharmaceuticals. Kushner is on the scientific advisory board for Novo Nordisk.
Marilynn Larkin, MA, is an award-winning medical writer and editor whose work has appeared in numerous publications, including Medscape Medical News and its sister publication MDedge, The Lancet (where she was a contributing editor), and Reuters Health.