BMI Isn’t Enough to Predict GLP-1 Risk. New Research Proves It.
A tool built from 200,000 people and published in Nature Medicine can now identify who will develop obesity complications before their BMI flags it. Here is what Obscore found — and what it means for GLP-1 eligibility thresholds.
BMI measures one thing: weight relative to height. It says nothing about visceral fat, insulin resistance, or glucose regulation. Those are the factors that actually drive disease risk. Obscore changes the eligibility conversation. For those already on GLP-1 therapy, none of this changes the nutrition strategy: protein first, muscle protected, deficit managed deliberately.
BMI Has Always Been a Blunt Instrument
BMI was developed in the 1830s as a statistical description of population averages. It was never a clinical tool for individual health. Adolphe Quetelet, the Belgian mathematician who created it, had no intention of it being used to decide who gets medical treatment. It measures mass relative to height squared and nothing else.
Two groups get routinely misclassified by BMI-based prescribing. The first: people with a BMI under 25 who carry significant visceral fat, show insulin resistance, and face genuine metabolic risk that their weight classification completely obscures. Research suggests this affects 20–30% of people in the normal BMI range. The second: people with a BMI over 30 whose metabolic markers are actually within normal ranges and whose real disease risk is considerably lower than their classification implies.
UK Biobank participants whose data was used to build and validate Obscore. One of the largest datasets ever used for obesity risk prediction.
Demircan et al., Nature Medicine, 2026Health factors analysed in the Obscore study, ranging from age and sex to complex blood biomarkers, to identify the 20 most predictive measures of obesity-related complications
Demircan et al., Nature Medicine, 2026Distinct obesity-related complications Obscore can predict, including type 2 diabetes, cancer, cardiovascular disease, and metabolic syndrome
Demircan et al., Nature Medicine, 2026What the Obscore Research Actually Found
The Obscore study, led by Professor Claudia Langenberg of Queen Mary University of London’s Precision Healthcare University Research Institute, analysed data from UK Biobank participants with a BMI of 27 and above. This was a deliberate choice to include the overweight range, not just those classified as obese. The researchers found that within the highest-risk group identified by Obscore, a considerable proportion of individuals were living with overweight rather than obesity.
As lead researcher Dr Kamil Demircan stated in the published findings: people in the overweight category who score high on Obscore face obesity-level metabolic complication risk despite not meeting current BMI thresholds for GLP-1 eligibility. These individuals, as the researchers note, constitute a population that may be overlooked if BMI alone is used as the decision framework.
The tool works by identifying 20 commonly collected health measures that already exist in most clinical records, and using them together to predict which individuals will develop complications. Professor Nick Wareham of the Institute of Metabolic Science at Cambridge, a co-author on the study, framed the challenge directly: GLP-1 therapies are now genuinely effective, but NHS resources are finite, which means allocation needs accurate and fair mechanisms. BMI alone does not provide that accuracy.
If you are already on Ozempic, Wegovy, Mounjaro, or Zepbound, Obscore does not change your treatment. It is a prescribing decision tool, not a nutrition framework. What it does confirm is something the metabolic health community has argued for years: your BMI number tells your clinician relatively little about your actual metabolic risk profile. The markers that matter (visceral fat, insulin sensitivity, glucose regulation) are the same markers that the Metabolic Nutrition System is built to address through structured fuelling and muscle preservation.
The Markers That Actually Determine Risk
Weight relative to height tells you nothing about what is happening inside the body. The factors that actually predict obesity complications are all about how the body processes and stores energy — not how much energy storage shows up on a scale.
Visceral Adiposity
Fat stored around internal organs (the liver, pancreas, kidneys, and intestines) is metabolically active in a way that subcutaneous fat is not. Visceral fat releases inflammatory cytokines, drives insulin resistance, and contributes directly to cardiovascular and metabolic disease risk. Two people with identical BMIs can have dramatically different visceral fat levels. That is why waist circumference and waist-to-hip ratio are significantly more predictive of metabolic risk than BMI. This is a central finding of metabolic age research and consistently appears in risk-stratification studies.
Insulin Resistance
Insulin resistance is the reduced ability of cells to respond to insulin signalling. It is the central mechanism in type 2 diabetes development and a driver of weight gain, visceral fat accumulation, and cardiovascular risk. It develops years before blood glucose rises to diabetic levels and is often invisible in standard BMI-based screening. GLP-1 medications work in part by enhancing insulin secretion and improving insulin sensitivity, which is why they are effective not just for weight loss but for metabolic disease management. A structured protein-first nutrition approach supports insulin sensitivity by reducing postprandial glucose spikes and maintaining lean mass, which is a key driver of insulin receptor density.
Blood Glucose Regulation
Fasting glucose and HbA1c (a marker of average blood glucose over the preceding 2–3 months) are more predictive of metabolic complication risk than BMI in multiple large studies. The relationship between glucose regulation and metabolic risk is linear. Risk worsens across the entire range from normal to prediabetes to diabetes, not only beyond a single threshold. People with a BMI of 27 and poor glucose regulation face greater complication risk than people with a BMI of 34 and normal glucose regulation. This is the population Obscore is specifically designed to identify. For GLP-1 users, adequate calorie intake is critical for glucose stability . Very low intake impairs the metabolic regulation the medication is working to improve.
Systemic Inflammation
Chronic low-grade inflammation, measured through markers such as C-reactive protein (CRP), interleukin-6, and tumour necrosis factor-alpha, is both a cause and consequence of metabolic dysfunction. Visceral fat drives inflammatory signalling. Inflammatory signalling in turn worsens insulin resistance, accelerates arterial damage, and increases cancer risk. These are three of the 18 complications Obscore is designed to predict. Nutrition plays a direct role: adequate protein intake supports the lean mass that regulates inflammatory pathways, while omega-3 fatty acids from fish, nuts, and seeds have direct anti-inflammatory effects supported by extensive trial data.
Lipid Profile
Elevated triglycerides, reduced HDL cholesterol, and the pattern of small dense LDL particles are all strongly predictive of cardiovascular metabolic risk, and all relatively independent of BMI. The triglyceride-to-HDL ratio in particular has emerged as a useful proxy for insulin resistance in clinical practice. A person with a BMI of 28 and a triglyceride-to-HDL ratio above 3.5 has significantly elevated cardiovascular risk that a BMI reading alone would not reveal. GLP-1 medications improve lipid profiles alongside weight loss, which is part of why the cardiovascular benefit of semaglutide and tirzepatide extends beyond what weight reduction alone would predict.
Current GLP-1 Eligibility: Where the Thresholds Stand Now
The prescribing thresholds vary by medication and by country. Here is where they currently stand — and where Obscore could shift them.
| Medication | NHS England (2026) | FDA (USA) | What Obscore may change |
|---|---|---|---|
| Semaglutide (Wegovy/Ozempic) | BMI ≥35 with weight-related condition, or BMI ≥40 | BMI ≥30, or ≥27 with weight-related condition | Could extend access to BMI 27–34 individuals with high Obscore metabolic risk |
| Tirzepatide (Mounjaro/Zepbound) | BMI ≥35 with weight-related health problems (NHS, phased rollout) | BMI ≥30, or ≥27 with weight-related condition | Tirzepatide criteria already includes health evidence beyond BMI — Obscore formalises this approach |
The distinction the Obscore researchers draw is important: tirzepatide’s NHS criteria already requires evidence of weight-related health problems alongside BMI. That is an early move toward the multi-marker approach Obscore proposes. The next step, as Professor Wareham described, is to incorporate a validated risk score so that clinicians can identify not just who has existing health problems, but who is at highest risk of developing them.
For current GLP-1 users, this development confirms the direction of travel: metabolic health markers, not weight classification alone, are becoming the central frame for managing obesity and its complications. This is the same frame that metabolic age research and the Metabolic Nutrition System have always operated within.
The Nutrition Strategy That Works at Any BMI
The metabolic markers Obscore prioritises (insulin resistance, glucose regulation, inflammation, visceral fat) are all responsive to the same nutrition and lifestyle interventions. BMI does not change what works. It changes the rate of progress and the degree of medication benefit, but the structural approach to nutrition remains constant.
Protein first, at every meal
Adequate protein intake supports insulin sensitivity through two mechanisms. First, it reduces the glycaemic impact of meals by slowing gastric emptying and reducing carbohydrate absorption rate. Second, it preserves and builds lean mass. Skeletal muscle is the largest site of insulin-mediated glucose uptake in the body. Every kilogram of lean mass lost during weight loss reduces the body’s capacity to manage blood glucose effectively. The muscle preservation research is unambiguous on this point: protecting lean mass is not cosmetic, it is metabolic.
Managing the calorie floor
Very low calorie intake is a common unintended outcome on GLP-1 medications. When appetite suppression pushes intake too low, it impairs glucose regulation rather than improving it. When intake falls below 800–900 calories per day, the body responds with adaptive thermogenesis and cortisol release, both of which worsen insulin sensitivity. The calorie deficit fatigue guide and the not eating enough on GLP-1 guide cover the warning signs and the practical floor management strategies that prevent this from becoming a problem.
Bone and metabolic density
Rapid weight loss, particularly at higher starting BMIs, accelerates bone mineral loss alongside fat loss. The Ozempic bone loss guide covers the mechanisms and the nutritional countermeasures: calcium, vitamin D, resistance training load, and adequate protein intake which supports bone remodelling through the same leucine-mTORC1 pathway that drives muscle protein synthesis.
The complete metabolic framework
The Metabolic Nutrition System is the central hub for the nutrition framework that addresses each of the metabolic markers Obscore prioritises. The GLP-1 Protein Calculator provides your personalised protein target. The Calorie Calculator helps establish a managed deficit that protects metabolic rate. For those on GLP-1 therapy managing adaptive thermogenesis, the structured fuelling approach prevents the metabolic adaptation that narrows the deficit over time.
Frequently Asked Questions
NHS England currently requires a BMI of 35 or above with at least one weight-related condition for semaglutide, or 35 plus evidence of weight-related health problems for tirzepatide. The FDA in the US covers BMI 30 or above, or 27 with a weight-related condition. These thresholds may evolve as risk-stratification tools like Obscore gain clinical adoption.
Obscore is a risk prediction tool published in Nature Medicine in 2026, built from data on over 200,000 UK Biobank participants and more than 2,000 health factors. It uses 20 commonly collected health measures to predict 18 different obesity-related complications, enabling identification of high-risk individuals in the overweight BMI range who are currently overlooked by BMI-only eligibility criteria.
Yes, this is well-established in metabolic medicine. Metabolically unhealthy normal weight affects approximately 20–30% of people with a BMI in the normal range. The critical factors are visceral fat, insulin resistance, blood glucose, inflammation, and lipid profiles. None of which BMI directly measures. Obscore addresses this gap specifically.
The strongest predictors of metabolic risk are: visceral adiposity (waist circumference, waist-to-hip ratio); insulin resistance (HOMA-IR, fasting insulin); blood glucose regulation (HbA1c, fasting glucose); systemic inflammation (CRP, IL-6); lipid profile (triglyceride-to-HDL ratio, LDL particle size); and liver function markers. These can be markedly abnormal in people with a healthy BMI, and surprisingly normal in people with a BMI over 35. That is exactly why BMI alone is insufficient for prescribing decisions.
The core principles remain constant: adequate protein (0.7–1.0g per pound of body weight), protein distributed across meals to hit the leucine threshold, resistance training as the muscle-preservation signal, and managed deficit that avoids extremes. What changes with BMI is primarily the rate of expected weight loss and the degree of appetite suppression on GLP-1 medications. The fundamental nutrition framework stays the same.
Research & References
- Demircan K, et al. Obscore: A data-driven tool for risk-stratification of obesity-related complications. Nature Medicine. 2026. nature.com
- Stefan N, et al. Metabolically healthy obesity: the low-risk obese phenotype and its relationship with cardiometabolic risk. Diabetologia. 2013;56(8):1622–1624. pubmed.ncbi.nlm.nih.gov
- Blüher M. Metabolically healthy obesity. Endocrine Reviews. 2020;41(3):405–420. pmc.ncbi.nlm.nih.gov
- Pontzer H, et al. Daily energy expenditure through the human life course. Science. 2021;373(6556):808–812. (Metabolic rate stability 20–60 — referenced in metabolic age research)
- Wilding JPH, et al. Once-weekly semaglutide in adults with overweight or obesity. New England Journal of Medicine. 2021;384:989–1002. (STEP 1 trial — cardiovascular and metabolic outcomes) pubmed.ncbi.nlm.nih.gov
- Jastreboff AM, et al. Tirzepatide once weekly for the treatment of obesity. New England Journal of Medicine. 2022;387:205–216. (SURMOUNT-1 trial) pubmed.ncbi.nlm.nih.gov
- NHS England. Tirzepatide (Mounjaro) criteria for specialist weight management services. Updated 2026.
- Science Advances 2024. UK Biobank N=225,212 — metabolic age and body composition analysis. (Referenced in metabolic age guide)