NEW YORK – Drugs targeting the glucagon-like peptide 1 (GLP-1) receptor have seized an outsized role in America's cardiometabolic health. By one estimate, more than half of all US adults are eligible for these drugs for diabetes or obesity, the two indications with US Food and Drug Administration approval. That's more than the portion of US adults eligible for statins. According to another estimate, nearly 12 percent of US adults have, at one point, taken a GLP-1 agonist.
The two biggest players in this market are Novo Nordisk and Eli Lilly. Novo Nordisk markets semaglutide under the brand name Ozempic for type 2 diabetes and as Wegovy for weight management, and Lilly sells tirzepatide for these same indications under the brand names Mounjaro and Zepbound, respectively.
Current market estimates likely understate the staggering use of this class of drugs, since providers often prescribe GLP-1 agonists off-label for conditions ranging from prediabetes and sleep apnea to polycystic ovary syndrome and substance use disorder. The soaring popularity of these drugs has translated into huge profits for the pharmaceutical companies behind them. One market research firm estimates that the global market for these drugs in 2024 is worth $47.4 billion and could grow to $471.1 billion by 2032, exceeding the value of the oncology drug market.
While the marketed GLP-1 receptor agonists differ slightly, the medications in this class generally work by simulating the effect of the naturally occurring hormone GLP-1, which regulates insulin, slows gastric emptying, and promotes satiety after eating. The drugs have helped millions of people, including high-profile celebrities, manage their diabetes and shed excess body weight.
As GLP-1 agonists continue to surge, however, researchers are beginning to question whether there's a better, more informed way to identify patients most likely to benefit. Because as famous and widely used as these drugs are, they don't work for everyone.
"If you look at the data, not everyone responds," said Andres Acosta, an associate professor of medicine at the Mayo Clinic and a cofounder of a biotech firm focused on obesity precision medicine called Phenomix Sciences. For example, with Wegovy for obesity, Acosta pointed to clinical data showing that after 64 weeks on this drug, only half of the patients lost 15 percent of their total body weight. And in real-world studies, only half of the patients on Wegovy lose more than 10 percent of their body weight after taking the drug for a year.
Diabetes endpoints have similarly varied responses. "The reality is that we need to find a good way to figure out who will respond to these medications," Acosta said.
GLP-1 agonists are also notoriously expensive, with list prices that amount to more than $1,000 per month. Plus, they carry the risk of gastrointestinal side effects such as nausea and vomiting. Given the varied benefits and side effects, a small but vocal group of researchers and diagnostic companies is beating the drum for developing precision treatment approaches for GLP-1 agonists.
"Right now, we're in the honeymoon phase where everyone is going on these drugs," said Quest Diagnostics Chief Medical Officer Yuri Fesko. "But there could be a future where we ask: 'Is this patient likely to respond to this drug before we start it?' We're not there yet, but I think that is going to happen."
Search for GLP-1 predictive biomarkers
The hunt for an effective precision medicine approach to guide GLP-1 receptor agonist use is still in its infancy, but researchers and diagnostic firms have started to lay the groundwork. Simeon Taylor, the director of the T32 Institutional Research Training Program in Diabetes and Obesity at the University of Maryland School of Medicine, for instance, is working to identify predictive genetic markers in the GLP receptor itself.
"If there were a scientific basis to predict with some level of confidence that this person's likely to respond and this person's not likely to respond, I think that could be useful," Taylor said. "But we're not there yet. There are some major hurdles."
Proving such biomarkers are useful for guiding treatment with these drugs will require extensive testing and clinical validation, he said. While such costly research can be partly funded with grants from the US National Institutes of Health, ultimately, it will require investment and buy-in from deep-pocketed drugmakers, who may be reluctant to explore biomarkers that can potentially limit the market size of these wildly in-demand drugs.
According to Taylor, much of the variation in patients' ability to respond to and tolerate GLP-1 agonists is related to behavioral or environmental differences. For example, given GLP-1 agonists' effect on the rate of gastric emptying, the process the stomach uses to shuttle its contents into the duodenum, patients who eat smaller meals more frequently while on the drugs are less likely to experience nausea than those who eat large meals in one sitting. Of course, the types of foods people eat, their activity levels, comorbidities, and a host of other factors play into the drugs' effectiveness in managing diabetes and obesity.
But genetics may play a role, too. With research funding from the NIH's National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), Taylor and his colleagues are actively conducting a genome-wide association study to explore the possibility that genetic variants may affect how well patients respond to GLP-1 receptor agonists for weight loss and diabetes.
Within an ongoing trial, these researchers are performing high-density array genotyping on 600 overweight or obese, but otherwise healthy, individuals who are part of the Old Order Amish community in Lancaster County, Pennsylvania. Taylor and his colleagues chose to conduct their research in this population because the community's shared lineage and homogenous lifestyle make it easier to isolate genes of interest. The researchers generated a global imputation panel for the genotyping portion of their study using whole-genome sequencing data derived from 1,025 Amish individuals.
At baseline, study participants receive an intravenous glucose tolerance (FSIGT) test, and then after taking Ozempic for six weeks, they receive another FSIGT test. Even though the volunteers in Taylor's study don't have diabetes, the researchers have shown through a separate pilot study that the FSIGT test can be a surrogate endpoint for how well Ozempic decreases blood glucose in diabetes patients.
In the genotyping portion of the study, Taylor and colleagues hope to identify variants to predict GLP-1 agonists' ability to accelerate the rate of glucose disappearance between the first and second FSIGT test.
It may be years before the researchers can point to any clear associations between these variants and outcomes garnered from the study. But Taylor, who worked in the drug industry for much of his career, knows firsthand that the work would be quicker, and more comprehensive, with pharma backing.
"We would be closer to that goal [of precision medicine for GLP-1 agonists] if industry and the FDA embraced this as something that was important to do," he said. "If companies would be willing to give their samples for DNA sequencing and let us leverage that data they've accumulated in a clinical trial to get the drug approved … it would be done less expensively."
Burgeoning biotech involvement
While the big drugmakers behind these GLP-1 agonists may not be spearheading efforts to develop GLP-1 precision medicine approaches, some biotechs and commercial labs are interested in exploring biomarker-informed strategies.
One health technology startup, Dandelion Health, has amassed a large patient-level dataset, which it markets for this type of research. Dandelion's GLP-1 Data Library includes multimodal real-world data from millions of patients, 200,000 of whom have received GLP-1 agonists. These data come from large, nonacademic US health systems, including Texas Health and the rural health system Sanford Health, among others.
"The patient population is incredibly diverse on purpose," said Dandelion CEO and Cofounder Elliott Green. "We're trying to encourage biotechs, pharma, artificial intelligence [tool] developers, and med device companies to use this data for the next evolution of creating biomarkers."
Although Dandelion's library doesn't include genomic data at the moment, Green said the firm may explore external partnerships to incorporate gene sequencing test results in the future. For now, the multimodal data include clinical information from electronic health records, radiology imaging, echocardiogram waveforms, and other clinical data from de-identified patients.
The firm is working with partners, including researchers and developers in the AI space, to use these data to advance algorithms for guiding GLP-1 treatment. "One person's clinical AI solution in the payor and provider space is another person's biomarker," Green said, underscoring that not all biomarkers are gleaned from gene-level data. Digital biomarkers as well as physiological and behavioral data, in Dandelion's view, are a promising area for precision medicine.
The firm is also conducting internal research using its data library to identify patients who don't have overt cardiovascular disease risk factors but might benefit from GLP-1 agonists as a preventive treatment. The firm recently put out a white paper detailing this research, though the team sees it as just the beginning of the various ways in which it can grow the GLP-1 data library.
"If you look at what happened in oncology using genomic data … there's now this incredible opportunity to do something somewhat similar using all of this other raw biological data that exists in the patient record that we've not touched before," Green said.
Phenotyping leads the way
Perhaps the most advanced attempt at applying precision medicine to GLP-1 receptor agonist treatment comes from Phenomix Sciences, the Mayo Clinic spinout that Acosta cofounded. Phenomix launched in 2021 after garnering exclusive rights to market the MyPhenome test from the Mayo Clinic, where it was developed. The company is selling MyPhenome as a tool to help select patients most likely to benefit from GLP-1 agonists for weight loss.
The saliva test is essentially a "version of a polygenic risk score," said Phenomix CEO Mark Bagnall. "We're looking at gene variants across 40 genes that we know have been demonstrated in the literature to have an effect on hunger and on obesity."
The test's algorithm factors in approximately 6,000 genetic variants across 40 genes, as well as patients' height, weight, and sex, and generates a score that can be used to stratify patients into four obesity phenotypes: Hungry Brain, Hungry Gut, Emotional Hunger, and Slow Burn.
Hungry Brain identifies those who consume too many calories without feeling full. Hungry Gut describes individuals who feel hungry shortly after eating. Emotional Hunger comprises patients who feel an emotional reward when they eat. People in the Slow Burn category are thought to have a defect in how they expend energy.
Acosta explained that the genes in the algorithm play a role in the ability of the gut-brain axis to regulate food intake. The GLP-1 receptor gene, which Maryland's Taylor and his colleagues are analyzing for predictive variants, is also among the genes in the MyPhenome risk score.
"We built a biomarker that can predict obesity phenotypes because we know that with the phenotypes, we can actually guide who should be on a GLP-1 or not," Acosta said. "After many years, we ended up with a gene risk score … to predict whether you're going to respond to this medication, or others."
The young company is still validating the MyPhenome test, but Acosta and Bagnall believe they have enough data showing that the test provides additional information that doctors can use to identify the right weight management treatments for their patients.
Earlier this year at the Digestive Disease Week conference, Acosta presented data on 84 patients with obesity who received Wegovy. Those who fell into the Hungry Gut phenotype, according to their MyPhenome genetic risk score, lost 19.5 percent of their total body weight after one year on Wegovy, whereas those with phenotypes other than Hungry Gut lost 10 percent of their body weight during the same period.
Acosta and colleagues have also produced data showing that obesity phenotyping can help guide treatment decisions across multiple classes of obesity drugs, including Novo Nordisk's earlier generation GLP-1 receptor agonist Saxenda (liraglutide).
In a study published in 2021, 84 patients who received obesity treatment via a phenotype-guided approach lost a mean 15.9 percent of their body weight after 12 months, whereas 228 patients who received obesity treatment without the phenotype-guided approach lost a mean 9 percent of their body weight in that time. Put another way, 79 percent of patients who received phenotype-guided treatment lost more than 10 percent of their total body weight after a year, whereas the same was true for just 34 percent of patients who received treatment without phenotyping.
Based on these studies, Acosta said Phenomix is seeing "a lot of early adopters" of their MyPhenome test among providers. Roughly 200 providers across the country are currently using this test, according to a Phenomix spokesperson.
"These are people who have looked at the data and evidence and say they believe in this and want to use it to help guide therapy," Acosta said. "But there are also people who say they would like to see some more data."
Angela O'Neil, a provider who treats patients with a range of health concerns at a Scottsdale, Arizona-based concierge medicine practice called Next Level Concierge Care, is one of Phenomix's early adopters. For about six months now, she has been using the MyPhenome test to help inform prescribing decisions for her patients who want to lose weight. She estimated ordering the test for 10 percent to 15 percent of her patients.
"This test should be a part of every metabolism program because it can help your patient understand their body and the best nutrition program as well as what adjunctive medication could help them lose weight," O'Neil said. "I now use this before I put any patient on a GLP-1 [agonist] because if you're not 'Hungry Gut,' it's not going to work."
O'Neil has seen at least one patient who started on a GLP-1 agonist but did not lose any weight and was later found to have a phenotype other than Hungry Gut based on the Phenomix test. But O'Neil recognizes that she may be a uniquely early adopter, given her private practice offers patients a suite of early healthtech innovations that may not be readily covered by insurance, from whole-body MRIs to multi-cancer early detection tests and even genetic tests that claim to screen for cognitive conditions.
Acosta acknowledged that other providers may still be skeptical or at least eager for more data. And, as of now, the test isn't covered by insurance.
He also recognized that the firm has only validated its test in people of European descent, so proving the test's performance in racial and ethnic minority patient populations will be key to broader adoption. "Our test was built on White Americans, so we need more diversity there," he said.
The firm also needs to assess the test's performance in prospective, large randomized-controlled trials as opposed to on real-world data. "We're just getting started," Acosta said, noting that studies in diverse patient populations as well as prospective studies are both underway.
In the meantime, Phenomix has partnered with the applied genomics solutions company InformedDNA to broaden MyPhenome's reach by tapping into relationships with payors and population health management programs.
"A lot of tests have suffered from a lack of rigor on the implementation and adoption side of the equation," said InformedDNA CEO Surya Singh. "The test is put out there, and we expect providers to learn about it, like the pharma model." Instead of sending out sales reps to educate doctors about their products like pharma companies do, a more effective strategy for a test like MyPhenome, in Singh's view, is to figure out how it fits into a program that a health plan sponsor is already implementing.
Through the recent collaboration, InformedDNA began offering Phenomix's MyPhenome test within its DNA Impact Metabolic program, aiming to make genetics-based weight management more accessible to partner organizations. Members and employees of the payors and life sciences organizations that have partnered with InformedDNA can get tested through the program and take the results to their providers for appropriate treatment and management, Singh said.
InformedDNA acknowledges in talks with partner organizations that MyPhenome is still early in its development cycle. But already, Singh has noticed an interest among payors that would welcome a way to lower costs by screening out patients who wouldn't benefit from GLP-1 receptor agonists.
"Payors are saying, 'If this can work, it would be a huge help to us because we have to wrestle with the decision of whether or not to cover GLP-1s for weight management,'" he said. As of now, he said, around 20 percent of self-funded employer health plans are covering GLP-1 receptor agonists for non-diabetes indications. And InformedDNA is hearing from these health plan providers that they want to be able to cover these drugs for these indications. They know their members are going out and getting these drugs anyway, sometimes through compounding pharmacies, and paying out of pocket. And while it's hard to prove, anecdotally Singh has heard of doctors prescribing GLP-1 receptor agonists for patients under their diabetes indications so that payors will cover them, even when weight loss is the main goal.
"These aren't, in my opinion, good or sustainable ways for GLP-1s to be managed," he said. But if plan sponsors, the majority of whom aren't currently covering these drugs for weight loss because of the sweeping cost they'd incur, can stratify patients most likely to benefit from these treatments, they might see a more cost-effective path toward this coverage.
Right now, the cost of the MyPhenome test is $499. The list price of Novo Nordisk's Wegovy is $1,349.02 per month, with competitor drugs priced in a similar range.
"If you do this [testing] upfront, a portion of the people who would have gotten GLP-1s will shift to another therapy or lifestyle management," Singh said, with the caveat that Phenomix will need to gather more data on its test showing that in can facilitate a cost-effective approach for administering these drugs.
Drugmakers in the driver's seat
Payors and providers may be enthusiastic about an approach that can identify patients most likely to response to GLP-1 drugs. But would the big pharmaceutical companies that market them welcome a test or technology that threatens to limit the market share of these top-selling products?
In 2023, Novo Nordisk's Ozempic and Wegovy brought in a combined $18.04 billion in sales, almost double the $9.36 billion in 2022. Eli Lilly's Zepbound was only approved in November 2023, but the drug still brought in $175.8 million in sales for that company in 2023, and Mounjaro brought in $5.16 billion.
The pharmaceutical "industry has defensible arguments for why [precision medicine approaches] maybe aren't the investments they want to make, at least in contrast to what they've done in cancer," Maryland's Taylor said, suggesting that limiting the pool of patients who receive these drugs isn't in the best interest of the companies that sell them. "What companies are concerned about sometimes is that if a precision medicine study shows that a drug doesn't work in half the people, they lose half of their market. That's not a great outcome for them."
In the cancer space, where drugmakers have embraced precision medicine approaches more readily, they've often done so only under regulatory pressure or after guidelines bodies have recommended doctors prescribe a drug to patients using biomarker testing based on available evidence. For example, AstraZeneca and Daiichi Sankyo recently withdrew a regulatory application for datopotamab deruxtecan in an all-comer non-small lung cancer population after discussing the available evidence with the FDA and submitted a new application seeking approval only in NSCLC patients with EGFR-mutated tumors.
In the absence of such incentives in the GLP-1 market, it's unlikely that drugmakers will be motivated to back precision medicine strategies that curtail the market for their drugs. InformedDNA's Singh predicted drug manufacturers that offer rebates to plan sponsors might refuse to provide them if the payor implements any sort of utilization management tool like a precision medicine test.
"Right now, the manufacturers are in the driver's seat saying: 'Do or don't do this or you don't get these rebates,'" Singh said. "That's a trade-off that the plan sponsors are having to make. They're having to look at whether the net economics are better for them. And if they're not, they have to say: 'This is something that I want to offer anyway because I don't want my people to be paying out of pocket for compounded semaglutide.'"
Lilly and Novo Nordisk didn't respond to questions about whether they're exploring precision medicine strategies for their GLP-1 agonist drugs before press time.
But according to Phenomix, the firm did just ink a partnership with Novo Nordisk, the maker of Ozempic and Wegovy. And while it's too soon for either company to share specifics about what the partnership will entail, one Novo Nordisk executive posted on LinkedIn that the collaboration will involve an in-depth study of people living with obesity to "explore whether [Novo Nordisk] can apply a precision medicine approach to obesity prevention and predict the predisposition to specific sub-groups of obesity."
"This partnership is just the start of our journey," Nadeem Sarwar, the head of Novo Nordisk's transformational prevention unit, wrote in his LinkedIn post.
"Novo Nordisk is very interested in understanding precision obesity," Acosta said. "They think there will be a future, in which patients stop this trial and error, and they believe in understanding the obesity's underlying pathology and treat it based on that."
Beyond the big pharma players, a precision medicine approach may be appealing to the dozens of drug companies developing their own GLP-1 candidates. "We've had discussions with half a dozen pharma companies," Bagnall said. "Their products aren't far enough along yet for discussion, but if they're running a Phase III trial and we can help them select patients that are more likely to be responders for that drug, they may want to identify those patients."
As companies look to develop truly differentiated obesity drugs, Bagnall suggested there might be a role for MyPhenome in identifying patients, for instance, those without the Hungry Gut phenotype, who would benefit from a new, non-GLP-1 agonist approach.
And ultimately, researchers working to validate new uses for GLP-1 agonists beyond diabetes and obesity say there's a definite role for precision medicine approaches. Lorenzo Leggio, the clinical director of the NIH's National Institute on Drug Abuse, is leading a large, randomized study to identify whether GLP-1 receptor agonists are effective in treating addiction.
"Once we finish our trial, in addition to answering the key question [as to whether] the drug worked compared to placebo, we also hope to answer additional questions," he said. "Who were the responders? Is there a specific phenotype? Are there biomarkers in the blood? It might not be a single biomarker, but there's no question that not everyone will respond to these medications, even if they do work."